Base_JSON/prefixN/json/N19/N19-1018.json

{
    "paper_id": "N19-1018",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T13:57:55.954464Z"
    },
    "title": "Discontinuous Constituency Parsing with a Stack-Free Transition System and a Dynamic Oracle",
    "authors": [
        {
            "first": "Maximin",
            "middle": [],
            "last": "Coavoux",
            "suffix": "",
            "affiliation": {},
            "email": "maximin.coavoux@naverlabs.com"
        },
        {
            "first": "Shay",
            "middle": [
                "B"
            ],
            "last": "Cohen",
            "suffix": "",
            "affiliation": {},
            "email": "scohen@inf.ed.ac.uk"
        }
    ],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "We introduce a novel transition system for discontinuous constituency parsing. Instead of storing subtrees in a stack-i.e. a data structure with linear-time sequential accessthe proposed system uses a set of parsing items, with constant-time random access. This change makes it possible to construct any discontinuous constituency tree in exactly 4n \u2212 2 transitions for a sentence of length n, whereas existing systems need a quadratic number of transitions to derive some structures. At each parsing step, the parser considers every item in the set to be combined with a focus item and to construct a new constituent in a bottomup fashion. The parsing strategy is based on the assumption that most syntactic structures can be parsed incrementally and that the setthe memory of the parser-remains reasonably small on average. Moreover, we introduce a dynamic oracle for the new transition system, and present the first experiments in discontinuous constituency parsing using a dynamic oracle. Our parser obtains state-of-the-art results on three English and German discontinuous treebanks.",
    "pdf_parse": {
        "paper_id": "N19-1018",
        "_pdf_hash": "",
        "abstract": [
            {
                "text": "We introduce a novel transition system for discontinuous constituency parsing. Instead of storing subtrees in a stack-i.e. a data structure with linear-time sequential accessthe proposed system uses a set of parsing items, with constant-time random access. This change makes it possible to construct any discontinuous constituency tree in exactly 4n \u2212 2 transitions for a sentence of length n, whereas existing systems need a quadratic number of transitions to derive some structures. At each parsing step, the parser considers every item in the set to be combined with a focus item and to construct a new constituent in a bottomup fashion. The parsing strategy is based on the assumption that most syntactic structures can be parsed incrementally and that the setthe memory of the parser-remains reasonably small on average. Moreover, we introduce a dynamic oracle for the new transition system, and present the first experiments in discontinuous constituency parsing using a dynamic oracle. Our parser obtains state-of-the-art results on three English and German discontinuous treebanks.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Abstract",
                "sec_num": null
            }
        ],
        "body_text": [
            {
                "text": "Discontinuous constituency trees extend standard constituency trees by allowing crossing branches to represent long distance dependencies, such as the wh-extraction in Figure 1 . Discontinuous constituency trees can be seen as derivations of Linear Context-Free Rewriting Systems (LCFRS, Vijay-Shanker et al., 1987) , a class of formal grammars more expressive than context-free grammars, which makes them much harder to parse. In particular, exact CKY-style LCFRS parsing has an O(n 3f ) time complexity where f is the fan-out of the grammar (Kallmeyer, 2010 A natural alternative to grammar-based chart parsing is transition-based parsing, that usually relies on fast approximate decoding methods such as greedy search or beam search. Transitionbased discontinuous parsers construct discontinuous constituents by reordering terminals with the SWAP action (Versley, 2014a,b; Maier, 2015; Maier and Lichte, 2016; Stanojevi\u0107 and Garrido Alhama, 2017) , or by using a split stack and the GAP action to combine two non-adjacent constituents (Coavoux and Crabb\u00e9, 2017a; Coavoux et al., 2019) . These proposals represent the memory of the parser (i.e. the tree fragments being constructed) with data structures with linear-time sequential access (either a stack, or a stack coupled with a double-ended queue). As a result, these systems need to perform at least n actions to construct a new constituent from two subtrees separated by n intervening subtrees. Our proposal aims at avoiding this cost when constructing discontinuous constituents.",
                "cite_spans": [
                    {
                        "start": 280,
                        "end": 315,
                        "text": "(LCFRS, Vijay-Shanker et al., 1987)",
                        "ref_id": null
                    },
                    {
                        "start": 543,
                        "end": 559,
                        "text": "(Kallmeyer, 2010",
                        "ref_id": "BIBREF29"
                    },
                    {
                        "start": 857,
                        "end": 875,
                        "text": "(Versley, 2014a,b;",
                        "ref_id": null
                    },
                    {
                        "start": 876,
                        "end": 888,
                        "text": "Maier, 2015;",
                        "ref_id": "BIBREF31"
                    },
                    {
                        "start": 889,
                        "end": 912,
                        "text": "Maier and Lichte, 2016;",
                        "ref_id": "BIBREF32"
                    },
                    {
                        "start": 913,
                        "end": 949,
                        "text": "Stanojevi\u0107 and Garrido Alhama, 2017)",
                        "ref_id": "BIBREF41"
                    },
                    {
                        "start": 1038,
                        "end": 1065,
                        "text": "(Coavoux and Crabb\u00e9, 2017a;",
                        "ref_id": "BIBREF6"
                    },
                    {
                        "start": 1066,
                        "end": 1087,
                        "text": "Coavoux et al., 2019)",
                        "ref_id": "BIBREF8"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 168,
                        "end": 176,
                        "text": "Figure 1",
                        "ref_id": "FIGREF0"
                    }
                ],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "We design a novel transition system in which a discontinuous constituent is constructed in a single step, without the use of reordering actions such as SWAP. The main innovation is that the memory of the parser is not represented by a stack, Table 1 : Set-based transition system description. Variable j is the number of steps performed since the start of the derivation.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 242,
                        "end": 249,
                        "text": "Table 1",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Buffer Stack",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "w i w i+1 w i+2 ... ... s 0 s 1 s 2 s 3 Buffer Set w i w i+1 w i+2 ... s 0 s 1 s n s f Focus ...",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "In a stack-based system like shift-reduceswap (upper part), the parser extracts features from a local region of the configuration (orange part), to predict the next action such as: construct a new tree with label X and children s 0 and s 1 (REDUCE-X). In our proposed set-based system (lower part), we consider every item in the set to be combined bottom-up with the focus item s f and score independently each possible transition.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Figure 2:",
                "sec_num": null
            },
            {
                "text": "as is usual in shift-reduce systems, but by an unordered random-access set. The parser considers every constituent in the current memory to construct a new constituent in a bottom-up fashion, and thus instantly models interactions between parsing items that are not adjacent. As such, we describe a left-to-right parsing model that deviates from the standard stack-buffer setting, a legacy from pushdown automata and classical parsing algorithms for context-free grammars.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Figure 2:",
                "sec_num": null
            },
            {
                "text": "Our contributions are summarized as follows:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Figure 2:",
                "sec_num": null
            },
            {
                "text": "\u2022 We design a novel transition system for discontinuous constituency parsing, based on a memory represented by a set of items, and that derives any tree in exactly 4n \u2212 2 steps for a sentence of length n; \u2022 we introduce the first dynamic oracle for discontinuous constituency parsing; \u2022 we present an empirical evaluation of the transition system and dynamic oracle on two German and one English discontinuous treebanks.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Figure 2:",
                "sec_num": null
            },
            {
                "text": "The code of our parser is released as an opensource project at https://gitlab.com/ mcoavoux/discoparset.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Figure 2:",
                "sec_num": null
            },
            {
                "text": "System overview We propose to represent the memory of the parser by (i) a set of parsing items and (ii) a single focus item. Figure 2 (lower part) illustrates a configuration in our system. The parser constructs a tree with two main actions: shift the next token to make it the new focus item (SHIFT), or combine any item in the set with the focus item to make a new constituent bottom-up (COMBINE action). Since the memory is not an ordered data structure, the parser considers equally every pending parsing item, and thus constructs a discontinuous constituent in a single step, thereby making it able to construct any discontinuous tree in O(n) transitions.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 125,
                        "end": 133,
                        "text": "Figure 2",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Set-based Transition System",
                "sec_num": "2"
            },
            {
                "text": "The use of an unordered random-access data structure to represent the memory of the parser also leads to a major change for the scoring system ( Figure 2 ). Stack-based systems use a local view of a parsing configuration to extract features and score actions: features only rely on the few topmost elements on the stack and buffer. The score of each transition depends on the totality of this local view. In constrast, we consider equally every item in the set, and therefore rely on a global Figure 1 . As a convention, we index elements in the set with their left-index and use COMB-i to denote COMB-s i . We also use tokens instead of their indexes for better legibility.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 145,
                        "end": 153,
                        "text": "Figure 2",
                        "ref_id": null
                    },
                    {
                        "start": 493,
                        "end": 501,
                        "text": "Figure 1",
                        "ref_id": "FIGREF0"
                    }
                ],
                "eq_spans": [],
                "section": "Set-based Transition System",
                "sec_num": "2"
            },
            {
                "text": "view of the memory (Section 3). However, we score each possible combinations independently: the score of a single combination only depends on the two constituents that are combined, regardless of the rest of the set.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Set-based Transition System",
                "sec_num": "2"
            },
            {
                "text": "Definitions We first define an instantiated (discontinuous) constituent (X, s) as a nonterminal label X associated with a set of token indexes s. We call min(s) the left-index of c and max(s) its right-index. For example in Figure 1 , the two VPs are respectively represented by (VP, {1, 6}) and (VP, {1, 5, 6}), and they have the same right index (6) and left index (1). A parsing configuration is a quadruple (S, s f , i, C) where:",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 224,
                        "end": 232,
                        "text": "Figure 1",
                        "ref_id": "FIGREF0"
                    }
                ],
                "eq_spans": [],
                "section": "System Description",
                "sec_num": "2.1"
            },
            {
                "text": "\u2022 S is a set of sets of indexes and represents the memory of the parser; \u2022 s f is a set of indexes called the focus item, and satisfies max(s f ) = i \u2212 1; \u2022 i is the index of the next token in the buffer;",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "System Description",
                "sec_num": "2.1"
            },
            {
                "text": "\u2022 C is a set of instantiated constituents.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "System Description",
                "sec_num": "2.1"
            },
            {
                "text": "Each new constituent is constructed bottom-up from the focus item and another item in the set S.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "System Description",
                "sec_num": "2.1"
            },
            {
                "text": "Transition set Our proposed transition system is based on the following types of actions:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "System Description",
                "sec_num": "2.1"
            },
            {
                "text": "\u2022 SHIFT constructs a singleton containing the next token in the buffer and assigns it as the new focus item. The former focus item is added to S.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "System Description",
                "sec_num": "2.1"
            },
            {
                "text": "\u2022 COMBINE-s computes the union between the focus item s f and another item s from the set S, to form the new focus item",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "System Description",
                "sec_num": "2.1"
            },
            {
                "text": "s \u222a s f . \u2022 LABEL-X instantiates a new constituent (X, s f )",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "System Description",
                "sec_num": "2.1"
            },
            {
                "text": "whose yield is the set of indexes in the focus item s f . \u2022 NO-LABEL has no effect; its semantics is that the current focus set is not a constituent.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "System Description",
                "sec_num": "2.1"
            },
            {
                "text": "Following Cross and Huang (2016b) , transitions are divided into structural actions (SHIFT, COM-BINE-s) and labelling actions (LABEL-X, NO-LABEL). The parser may only perform a structural action on an even step and a labelling action on an odd step. For our system, this distinction has the crucial advantage of keeping the number of possible actions low at each parsing step, compared to a system that would perform a COMBINE action and a labelling action in a single REDUCE-s-X action. 1 Table 1 presents each action as a deduction rule associated with preconditions. In Table 2 , we describe how to derive the tree from Figure 1.",
                "cite_spans": [
                    {
                        "start": 10,
                        "end": 33,
                        "text": "Cross and Huang (2016b)",
                        "ref_id": "BIBREF14"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 490,
                        "end": 497,
                        "text": "Table 1",
                        "ref_id": null
                    },
                    {
                        "start": 573,
                        "end": 580,
                        "text": "Table 2",
                        "ref_id": "TABREF2"
                    }
                ],
                "eq_spans": [],
                "section": "System Description",
                "sec_num": "2.1"
            },
            {
                "text": "Training a transition-based parser requires an oracle, i.e. a function that determines what the best action is in a specific parsing configuration to serve as a training signal. We first describe a static oracle that provides a canonical derivation for a given gold tree. We then introduce a dynamic oracle that determines what the best action is in any parsing configuration.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Oracles",
                "sec_num": "2.2"
            },
            {
                "text": "Our transition system exhibits a fair amount of spurious ambiguity, the ambiguity exhibited by the existence of many possible derivations for a single tree. Indeed, since we use an unordered memory, an n-ary constituent (and more generally a tree) can be constructed by many different transition sequences. For example, the set {0, 1, 2} might be constructed by combining",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Static Oracle",
                "sec_num": "2.2.1"
            },
            {
                "text": "\u2022 {0} and {1} first, and the result with {2}; or \u2022 {1} and {2} first, and the result with {0}; or \u2022 {0} and {2} first, and the result with {1}.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Static Oracle",
                "sec_num": "2.2.1"
            },
            {
                "text": "Following Cohen et al. 2012, we eliminate spurious ambiguity by selecting a canonical derivation for a gold tree. In particular, we design the static oracle (i) to apply COMBINE as soon as possible in order to minimize the size of the memory (ii) to combine preferably with the most recent set in the memory when several combinations are possible. The first choice is motivated by properties of our system: when the memory is smaller, there are fewer choices, therefore decisions are simpler and less expensive to score.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Static Oracle",
                "sec_num": "2.2.1"
            },
            {
                "text": "Parsers are usually trained to predict the gold sequence of actions, using a static oracle. The limitation of this method is that the parser only sees a tiny portion of the search space at train time and only trains on gold input (i.e. configurations obtained after performing gold actions). At test time, it is in a different situation due to error propagation: it must predict what the best actions are in configurations from which the gold tree is probably no longer reachable.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "To alleviate this limitation, Goldberg and Nivre (2012) proposed to train a parser with a dynamic oracle, an oracle that is defined for any parsing configuration and outputs the set of best actions to perform. In contrast, a static oracle is deterministic and is only defined for gold configurations.",
                "cite_spans": [
                    {
                        "start": 30,
                        "end": 55,
                        "text": "Goldberg and Nivre (2012)",
                        "ref_id": "BIBREF24"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "Dynamic oracles were proposed for a wide range of dependency parsing transition systems G\u00f3mez-Rodr\u00edguez et al., 2014; G\u00f3mez-Rodr\u00edguez and Fern\u00e1ndez-Gonz\u00e1lez, 2015) , and later adapted to constituency parsing (Coavoux and Crabb\u00e9, 2016; Huang, 2016b; Fern\u00e1ndez-Gonz\u00e1lez and G\u00f3mez-Rodr\u00edguez, 2018b,a).",
                "cite_spans": [
                    {
                        "start": 88,
                        "end": 117,
                        "text": "G\u00f3mez-Rodr\u00edguez et al., 2014;",
                        "ref_id": "BIBREF27"
                    },
                    {
                        "start": 118,
                        "end": 163,
                        "text": "G\u00f3mez-Rodr\u00edguez and Fern\u00e1ndez-Gonz\u00e1lez, 2015)",
                        "ref_id": "BIBREF26"
                    },
                    {
                        "start": 208,
                        "end": 234,
                        "text": "(Coavoux and Crabb\u00e9, 2016;",
                        "ref_id": "BIBREF5"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "In the remainder of this section, we introduce a dynamic oracle for our proposed transition system. It can be seen as an extension of the oracle of Cross and Huang (2016b) to the case of discontinuous parsing.",
                "cite_spans": [
                    {
                        "start": 148,
                        "end": 171,
                        "text": "Cross and Huang (2016b)",
                        "ref_id": "BIBREF14"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "Preliminary definitions For a parsing configuration c, the relation c c holds iff c can be derived from c by a single transition. We note * the reflexive and transitive closure of . An instantiated constituent (X, s) is reachable from a configuration c = (S, s f , i, C) iff there exists c = (S , s f , i , C ) such that (X, s) \u2208 C and c * c . Similarly, a set of constituents t (possibly a full discontinuous constituency tree) is reachable iff there exists a configuration c = (S , s f , i , C ) such that t \u2286 C and c * c . We note reach(c, t * ) the set of constituents that are (i) in the gold set of constituents t * (ii) reachable from c.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "We define a total order on index sets:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "s s \u21d4 \uf8f1 \uf8f4 \uf8f4 \uf8f2 \uf8f4 \uf8f4 \uf8f3 max(s) < max(s ), or max(s) = max(s ) and s \u2286 s .",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "This order naturally extends to the constituents of a tree:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "(X, s) (X , s ) iff s s . If (X, s) precedes (X , s )",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": ", then (X, s) must be constructed before (X , s ). Indeed, since the right-index of the focus item is non-decreasing during a derivation (as per the transition definitions), constituents are constructed in the order of their right-index (first condition). Moreover, since the algorithm is bottom-up, a constituent must be constructed before its parent (second condition).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "From a configuration c = (S, s f , i, C) at an odd step, a constituent (X, s g ) / \u2208 C is reachable iff both the following properties hold:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "1. max(s f ) \u2264 max(s g ); 2. \u2200s \u2208 S \u222a {s f }, (s \u2286 s g ) or (s \u2229 s g = \u2205).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "Condition 1 is necessary because the parser can only construct new constituents (X, s) such that s f s. Condition 2 makes sure that s g can be constructed from a union of elements from S \u222a {s f }, potentially augmented with terminals from the bufffer: {i, i + 1, . . . , max(s g )}.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "Following Cross and Huang (2016b), we define next(c, t * ) as the smallest reachable gold constituent from a configuration c. Formally:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "next(c, t * ) = argmin reach(c, t * ).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "Oracle algorithm We first define the oracle o for the odd step of a configuration c = (S, s f , i, C):",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "o odd (c, t * ) = {LABEL-X} if \u2203(X, s f ) \u2208 t * , {NO-LABEL} otherwise.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "For even steps, assuming next(c, t * ) = (X, s g ), we define the oracle as follows:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "o even (c, t * ) = \uf8f1 \uf8f4 \uf8f4 \uf8f2 \uf8f4 \uf8f4 \uf8f3 {COMB-s|(s f \u222a s) \u2286 s g } if max(s g ) = max(s f ), {COMB-s|(s f \u222a s) \u2286 s g } \u222a {SH} if max(s g ) > max(s f ).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "We provide a proof of the correctness of the oracle in Appendix A.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Dynamic Oracle",
                "sec_num": "2.2.2"
            },
            {
                "text": "We first present an encoder that computes contextaware representations of tokens (Section 3.1). We then discuss how to compute the representation of a set of tokens (Section 3.2). We describe the action scorer (Section 3.3), the POS tagging component (Section 3.4), and the objective function (Section 3.5).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Neural Network based on Constituent Boundaries",
                "sec_num": "3"
            },
            {
                "text": "As in recent proposals in dependency and constituency parsing (Cross and Huang, 2016a; Kiperwasser and Goldberg, 2016), our scoring system is based on a sentence transducer that constructs a context-aware representation for each token. Given a sequence of tokens x n 1 = (x 1 , . . . , x n ), we first run a single-layer character bi-LSTM encoder c to obtain a character-aware embedding c(x i ) for each token. We represent a token x i as the concatenation of a standard word embedding e(x i ) and the character-aware embedding:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Token Representations",
                "sec_num": "3.1"
            },
            {
                "text": "w x i = [c(x i ); e(x i )].",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Token Representations",
                "sec_num": "3.1"
            },
            {
                "text": "Then, we run a 2-layer bi-LSTM transducer over the sequence of token representations:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Token Representations",
                "sec_num": "3.1"
            },
            {
                "text": "(h (1) 1 , . . . , h (1) n ) = bi-LSTM(w x 1 , . . . , w xn ), (h (2) 1 , . . . , h (2) n ) = bi-LSTM(h (1) 1 , . . . , h (1) n ).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Token Representations",
                "sec_num": "3.1"
            },
            {
                "text": "The parser uses the context-aware token representations h",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Token Representations",
                "sec_num": "3.1"
            },
            {
                "text": "i to construct vector representations of sets or constituents.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Token Representations",
                "sec_num": "3.1"
            },
            {
                "text": "An open issue in neural discontinuous parsing is the representation of discontinuous constituents. In projective constituency parsing, it has become standard to use the boundaries of constituents (Hall et al., 2014; Crabb\u00e9, 2015; Durrett and Klein, 2015) , an approach that proved very successful with bi-LSTM token representations (Cross and Huang, 2016b; Stern et al., 2017) .",
                "cite_spans": [
                    {
                        "start": 196,
                        "end": 215,
                        "text": "(Hall et al., 2014;",
                        "ref_id": "BIBREF28"
                    },
                    {
                        "start": 216,
                        "end": 229,
                        "text": "Crabb\u00e9, 2015;",
                        "ref_id": "BIBREF11"
                    },
                    {
                        "start": 230,
                        "end": 254,
                        "text": "Durrett and Klein, 2015)",
                        "ref_id": "BIBREF16"
                    },
                    {
                        "start": 332,
                        "end": 356,
                        "text": "(Cross and Huang, 2016b;",
                        "ref_id": "BIBREF14"
                    },
                    {
                        "start": 357,
                        "end": 376,
                        "text": "Stern et al., 2017)",
                        "ref_id": "BIBREF42"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Set Representations",
                "sec_num": "3.2"
            },
            {
                "text": "Although constituent boundary features improves discontinuous parsing (Coavoux and Crabb\u00e9, 2017a) , relying only on the left-index and the right-index of a constituent has the limitation of ignoring gaps inside a constituent. For example, since the two VPs in Figure 1 have the same right-index and left-index, they would have the same representations. It may also happen that constituents with identical right-index and left-index do not have the same labels.",
                "cite_spans": [
                    {
                        "start": 70,
                        "end": 97,
                        "text": "(Coavoux and Crabb\u00e9, 2017a)",
                        "ref_id": "BIBREF6"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 260,
                        "end": 268,
                        "text": "Figure 1",
                        "ref_id": "FIGREF0"
                    }
                ],
                "eq_spans": [],
                "section": "Set Representations",
                "sec_num": "3.2"
            },
            {
                "text": "We represent a (possibly partial) constituent with the yield s, by computing 4 indexes from s: (min(s), max(s), min(s), max(s)). The set s represents the gap in s, i.e. the tokens between min(s) and max(s) that are not in the yield of s:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Set Representations",
                "sec_num": "3.2"
            },
            {
                "text": "s = {i| min(s) < i < max(s) and i / \u2208 s}.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Set Representations",
                "sec_num": "3.2"
            },
            {
                "text": "Finally, we extract the corresponding token representations of the 4 indexes and concatenate them to form the vector representation r(s) of s:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Set Representations",
                "sec_num": "3.2"
            },
            {
                "text": "r(s) = [h (2) min(s) ; h (2) max(s) ; h (2) min(s) ; h (2) max(s) ].",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Set Representations",
                "sec_num": "3.2"
            },
            {
                "text": "For an index set that does not contain a gap, we have s = \u2205. To handle this case, we use a parameter vector h nil , randomly initialized and learned jointly with the network, to embed max(\u2205) = min(\u2205) = nil. For example, the constituents (VP, {1, 6}) and (VP, {1, 5, 6}) will be respectively vectorized as:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Set Representations",
                "sec_num": "3.2"
            },
            {
                "text": "r({1, 6}) = [h (2) 1 ; h (2) 6 ; h (2) 2 ; h (2) 5 ], r({1, 5, 6}) = [h (2) 1 ; h (2) 6 ; h (2) 2 ; h (2) 4 ].",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Set Representations",
                "sec_num": "3.2"
            },
            {
                "text": "This representation method makes sure that two distinct index sets have distinct representations, as long as they have at most one gap each. This property no longer holds if one index sets has more than one gap.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Set Representations",
                "sec_num": "3.2"
            },
            {
                "text": "For each type of action -structural or labelling-we use a feedforward network with two hidden layers.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Action Scorer",
                "sec_num": "3.3"
            },
            {
                "text": "Structural actions At structural steps, for a configuration c = (S, s f , i, C), we need to compute the score of |S| COMBINE actions and possibly a SHIFT action. In our approach, the score of a combine-s action only depends on s and s f and is independent of the rest of the configuration (i.e. other items in the set). We first construct input matrix M as follows:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Action Scorer",
                "sec_num": "3.3"
            },
            {
                "text": "M = r(s 1 ) \u2022 \u2022 \u2022 r(s n ) r({i}) r(s f ) \u2022 \u2022 \u2022 r(s f ) r(s f ) .",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Action Scorer",
                "sec_num": "3.3"
            },
            {
                "text": "Each of the first n columns of matrix M represents the input for a COMBINE action, whereas the last column is the input for the SHIFT action. We then compute the score of each structural action:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Action Scorer",
                "sec_num": "3.3"
            },
            {
                "text": "P (\u2022|c) = Softmax(FF s (M )),",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Action Scorer",
                "sec_num": "3.3"
            },
            {
                "text": "where FF s is a feedforward network with two hidden layers, a tanh activation and a single output unit. In other words, it outputs a single scalar for each column vector of matrix M . This part of the network can be seen as an attention mechanism, where the focus item is the query, and the context is formed by the items in the set and the first element in the buffer.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Action Scorer",
                "sec_num": "3.3"
            },
            {
                "text": "We compute the probabilities of labelling actions as follows:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Labelling actions",
                "sec_num": null
            },
            {
                "text": "P (\u2022|s f ) = Softmax(FF l (r(s f ))),",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Labelling actions",
                "sec_num": null
            },
            {
                "text": "where FF l is a feedforward network with two hidden layers activated with the tanh function, and |N | + 1 output units, where N is the set of nonterminals.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Labelling actions",
                "sec_num": null
            },
            {
                "text": "Following Coavoux and Crabb\u00e9 (2017b), we use the first layer of the bi-LSTM transducer as input to a Part-of-Speech (POS) tagger that is learned jointly with the parser. For a sentence x n 1 , we compute the probability of a sequence of POS tags t n 1 = (t 1 , . . . , t n ) as follows:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "POS Tagger",
                "sec_num": "3.4"
            },
            {
                "text": "P (t n 1 |x n 1 ) = n i=1 Softmax(W (t) \u2022 h (1) i + b (t) ) t i ,",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "POS Tagger",
                "sec_num": "3.4"
            },
            {
                "text": "where W (t) and b (t) are parameters.",
                "cite_spans": [
                    {
                        "start": 8,
                        "end": 11,
                        "text": "(t)",
                        "ref_id": null
                    },
                    {
                        "start": 18,
                        "end": 21,
                        "text": "(t)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "POS Tagger",
                "sec_num": "3.4"
            },
            {
                "text": "In the static oracle setting, for a single sentence x n 1 , we optimize the sum of the loglikelihood of gold POS-tags t n 1 and the loglikelihood of gold parsing actions a n 1 :",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Objective Function",
                "sec_num": "3.5"
            },
            {
                "text": "L = L t + L p , L t = \u2212 n i=1 log P (t i |x n 1 ), L p = \u2212 4n\u22122 i=1 log P (a i |a i\u22121 1 , x n 1 ).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Objective Function",
                "sec_num": "3.5"
            },
            {
                "text": "We optimize this objective by alternating a stochastic step for the tagging objective and a stochastic step for the parsing objective, as is standard in multitask learning (Caruana, 1997) .",
                "cite_spans": [
                    {
                        "start": 172,
                        "end": 187,
                        "text": "(Caruana, 1997)",
                        "ref_id": "BIBREF4"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Objective Function",
                "sec_num": "3.5"
            },
            {
                "text": "In the dynamic oracle setting, instead of optimizing the likelihood of the gold actions (assuming all previous actions were gold), we optimize the likelihood of the best actions, as computed by the dynamic oracle, from a configuration sampled from the space of all possible configurations. In practice, before each epoch, we sample each sentence from the training corpus with probability p and we use the current (non-averaged) parameters to parse the sentence and generate a sequence of configurations. Instead of selecting the highestscoring action at each parsing step, as in a normal inference step, we sample an action using the softmax distribution computed by the parser, as done by Ballesteros et al. (2016) . Then, we use the dynamic oracle to calculate the best action from each of these configurations. In case there are several best actions, we deterministically choose a single action by favoring a COMBINE over a SHIFT (to bias the model towards a small memory), and to COMBINE with the item with the highest right-index (to avoid spurious discontinuity in partial constituents). We train the parser on these sequences of potentially non-gold configurationaction pairs.",
                "cite_spans": [
                    {
                        "start": 690,
                        "end": 715,
                        "text": "Ballesteros et al. (2016)",
                        "ref_id": "BIBREF1"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Objective Function",
                "sec_num": "3.5"
            },
            {
                "text": "We carried out experiments to assess the adequacy of our system and the effect of training with the dynamic oracle. We present the three discontinuous constituency treebanks that we used (Section 4.1), our experimental protocol (Section 4.2), then we discuss the results (Section 4.3) and the efficiency of the parser (Section 4.4).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments",
                "sec_num": "4"
            },
            {
                "text": "Tiger ",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "DPTB",
                "sec_num": null
            },
            {
                "text": "Negra F1",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "DPTB",
                "sec_num": null
            },
            {
                "text": "We perform experiments on three discontinuous constituency corpora. The discontinuous Penn Treebank was introduced by Evang and Kallmeyer (2011) who converted the long distance dependencies encoded by indexed traces in the original Penn treebank (Marcus et al., 1993) to discontinuous constituents. We used the standard split (sections 2-21 for training, 22 for development and 23 for test). The Tiger corpus (Brants et al., 2004) and the Negra corpus (Skut et al., 1997) are both German treebanks natively annotated with discontinuous constituents. We used the SPMRL split for the Tiger corpus (Seddah et al., 2013) , and the split of Dubey and Keller (2003) for the Negra corpus.",
                "cite_spans": [
                    {
                        "start": 246,
                        "end": 267,
                        "text": "(Marcus et al., 1993)",
                        "ref_id": "BIBREF33"
                    },
                    {
                        "start": 409,
                        "end": 430,
                        "text": "(Brants et al., 2004)",
                        "ref_id": "BIBREF3"
                    },
                    {
                        "start": 452,
                        "end": 471,
                        "text": "(Skut et al., 1997)",
                        "ref_id": "BIBREF40"
                    },
                    {
                        "start": 595,
                        "end": 616,
                        "text": "(Seddah et al., 2013)",
                        "ref_id": "BIBREF39"
                    },
                    {
                        "start": 636,
                        "end": 659,
                        "text": "Dubey and Keller (2003)",
                        "ref_id": "BIBREF15"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Datasets",
                "sec_num": "4.1"
            },
            {
                "text": "We implemented our parser in Python using the Pytorch library (Paszke et al., 2017) . We trained each model with the ASGD algorithm (Polyak and Juditsky, 1992) for 100 epochs. Training a single model takes approximately a week with a GPU. We evaluate a model every 4 epochs on the validation set and select the best performing model according to the validation F-score. We refer the reader to Table 5 of Appendix B for the full list of hyperparameters.",
                "cite_spans": [
                    {
                        "start": 62,
                        "end": 83,
                        "text": "(Paszke et al., 2017)",
                        "ref_id": "BIBREF36"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 393,
                        "end": 400,
                        "text": "Table 5",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Implementation and Protocol",
                "sec_num": "4.2"
            },
            {
                "text": "We evaluate models with the dedicated module of discodop 2 (van Cranenburgh et al., 2016) . We use the standard evaluation parameters (proper.prm), that ignore punctuations and root symbols. We report two evaluation metrics: a standard Fscore (F) and an Fscore computed only on discontinuous constituents (Disc. F), which provides a more qualitative evaluation of the ability of the parser to recover long distance dependencies.",
                "cite_spans": [
                    {
                        "start": 59,
                        "end": 89,
                        "text": "(van Cranenburgh et al., 2016)",
                        "ref_id": "BIBREF12"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Implementation and Protocol",
                "sec_num": "4.2"
            },
            {
                "text": "We present parsing results on the development sets of each corpus in Table 3 . The effect of the oracle is in line with other published results in projective constituency parsing (Coavoux and Crabb\u00e9, 2016; Cross and Huang, 2016b) and dependency parsing (Goldberg and Nivre, 2012; G\u00f3mez-Rodr\u00edguez et al., 2014) : the dynamic oracle improves the generalization capability of the parser. Table 4 , we compare our parser to other transition-based parsers (Maier, 2015; Coavoux and Crabb\u00e9, 2017a; Stanojevi\u0107 and Garrido Alhama, 2017; Coavoux et al., 2019) , the pseudo-projective parser of Versley (2016) , grammar-based chart parsers (Evang and Kallmeyer, 2011; van Cranenburgh et al., 2016; Gebhardt, 2018) and parsers based on dependency parsing (Fern\u00e1ndez-Gonz\u00e1lez and Martins, 2015; Corro et al., 2017) . Note that some of them only report results in a gold POS tag setting (the parser has access to gold POS tags and use them as features), a setting that is much easier than ours.",
                "cite_spans": [
                    {
                        "start": 179,
                        "end": 205,
                        "text": "(Coavoux and Crabb\u00e9, 2016;",
                        "ref_id": "BIBREF5"
                    },
                    {
                        "start": 206,
                        "end": 229,
                        "text": "Cross and Huang, 2016b)",
                        "ref_id": "BIBREF14"
                    },
                    {
                        "start": 253,
                        "end": 279,
                        "text": "(Goldberg and Nivre, 2012;",
                        "ref_id": "BIBREF24"
                    },
                    {
                        "start": 280,
                        "end": 309,
                        "text": "G\u00f3mez-Rodr\u00edguez et al., 2014)",
                        "ref_id": "BIBREF27"
                    },
                    {
                        "start": 451,
                        "end": 464,
                        "text": "(Maier, 2015;",
                        "ref_id": "BIBREF31"
                    },
                    {
                        "start": 465,
                        "end": 491,
                        "text": "Coavoux and Crabb\u00e9, 2017a;",
                        "ref_id": "BIBREF6"
                    },
                    {
                        "start": 492,
                        "end": 528,
                        "text": "Stanojevi\u0107 and Garrido Alhama, 2017;",
                        "ref_id": "BIBREF41"
                    },
                    {
                        "start": 529,
                        "end": 550,
                        "text": "Coavoux et al., 2019)",
                        "ref_id": "BIBREF8"
                    },
                    {
                        "start": 585,
                        "end": 599,
                        "text": "Versley (2016)",
                        "ref_id": "BIBREF45"
                    },
                    {
                        "start": 630,
                        "end": 657,
                        "text": "(Evang and Kallmeyer, 2011;",
                        "ref_id": "BIBREF17"
                    },
                    {
                        "start": 658,
                        "end": 687,
                        "text": "van Cranenburgh et al., 2016;",
                        "ref_id": "BIBREF12"
                    },
                    {
                        "start": 688,
                        "end": 703,
                        "text": "Gebhardt, 2018)",
                        "ref_id": "BIBREF22"
                    },
                    {
                        "start": 744,
                        "end": 782,
                        "text": "(Fern\u00e1ndez-Gonz\u00e1lez and Martins, 2015;",
                        "ref_id": "BIBREF21"
                    },
                    {
                        "start": 783,
                        "end": 802,
                        "text": "Corro et al., 2017)",
                        "ref_id": "BIBREF10"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 69,
                        "end": 76,
                        "text": "Table 3",
                        "ref_id": "TABREF4"
                    },
                    {
                        "start": 385,
                        "end": 392,
                        "text": "Table 4",
                        "ref_id": "TABREF5"
                    }
                ],
                "eq_spans": [],
                "section": "Effect of Dynamic Oracle",
                "sec_num": null
            },
            {
                "text": "Our parser matches the state of the art of Coavoux et al. (2019) . This promising result shows that it is feasible to design accurate transition systems without an ordered memory.",
                "cite_spans": [
                    {
                        "start": 43,
                        "end": 64,
                        "text": "Coavoux et al. (2019)",
                        "ref_id": "BIBREF8"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "External comparisons In",
                "sec_num": null
            },
            {
                "text": "Our transition system derives a tree for a sentence of n words in exactly 4n \u2212 2 transitions. Indeed, there must be n SHIFT actions, and n\u22121 COMBINE actions. Each of these 2n \u2212 1 transitions must be followed by a single labelling action.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Efficiency",
                "sec_num": "4.4"
            },
            {
                "text": "The statistical model responsible for choosing which action to perform at each parsing step needs to score |S| + 1 actions for a structural step and |N |+1 actions for a labelling step (where N is the set of possible nonterminals). Since in the worst case, |S| contains n \u2212 1 singletons, the parser has an O(n(|N | + n)) time complexity.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Efficiency",
                "sec_num": "4.4"
            },
            {
                "text": "In practice, the memory of the parser S remains relatively small on average. We report in Figure 3 the distribution of the size of S across configurations when parsing the development sets of three corpora. For the German treebanks, the memory contains 7 or fewer elements for more than 99 percents of configurations. For the Penn treebank, the memory is slighlty larger, with 98 percents of configuration with 11 or fewer items. Figure 3 : Distribution of the size of the memory of the parser S across configurations derived when parsing the development set of the three corpora. In practice, we observe that the memory remains small.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 90,
                        "end": 98,
                        "text": "Figure 3",
                        "ref_id": null
                    },
                    {
                        "start": 430,
                        "end": 438,
                        "text": "Figure 3",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Efficiency",
                "sec_num": "4.4"
            },
            {
                "text": "We report empirical runtimes in Table 6 of Appendix C. Our parser compares decently with other transition-based parsers, despite being written in Python.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 32,
                        "end": 39,
                        "text": "Table 6",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Efficiency",
                "sec_num": "4.4"
            },
            {
                "text": "Existing transition systems for discontinuous constituency parsing rely on three main strategies for constructing discontinuous constituents: a swap-based strategy, a split-stack strategy, and the use of non-local transitions.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "5"
            },
            {
                "text": "Swap-based systems Swap-based transition systems are based on the idea that any discontinuous constituency tree can be transformed into a projective tree by reordering terminals. They reorder terminals by swapping them with a dedicated action (SWAP), commonly used in dependency parsing (Nivre, 2009) . The first proposals in transition-based discontinuous constituency parsing used the SWAP action on top of an easy-first parser (Versley, 2014a,b) . Subsequent proposals relied on a shift-reduce system (Maier, 2015; Maier and Lichte, 2016) or a shift-promote-adjoin system (Stanojevi\u0107 and Garrido Alhama, 2017) .",
                "cite_spans": [
                    {
                        "start": 287,
                        "end": 300,
                        "text": "(Nivre, 2009)",
                        "ref_id": "BIBREF35"
                    },
                    {
                        "start": 430,
                        "end": 448,
                        "text": "(Versley, 2014a,b)",
                        "ref_id": null
                    },
                    {
                        "start": 504,
                        "end": 517,
                        "text": "(Maier, 2015;",
                        "ref_id": "BIBREF31"
                    },
                    {
                        "start": 518,
                        "end": 541,
                        "text": "Maier and Lichte, 2016)",
                        "ref_id": "BIBREF32"
                    },
                    {
                        "start": 575,
                        "end": 612,
                        "text": "(Stanojevi\u0107 and Garrido Alhama, 2017)",
                        "ref_id": "BIBREF41"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "5"
            },
            {
                "text": "The main limitation of swap-based system is that they tend to require a large number of transitions to derive certain trees. The choice of an oracle that minimizes derivation lengths has a substantially positive effect on parsing (Maier and Lichte, 2016; Stanojevi\u0107 and Garrido Alhama, 2017) .",
                "cite_spans": [
                    {
                        "start": 230,
                        "end": 254,
                        "text": "(Maier and Lichte, 2016;",
                        "ref_id": "BIBREF32"
                    },
                    {
                        "start": 255,
                        "end": 291,
                        "text": "Stanojevi\u0107 and Garrido Alhama, 2017)",
                        "ref_id": "BIBREF41"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "5"
            },
            {
                "text": "Split-stack systems The second parsing strategy constructs discontinuous constituents by allowing the parser to reduce pairs of items that are not adjacent in the stack. In practice, Coavoux and Crabb\u00e9 (2017a) split the usual stack of shiftreduce parsers into two data structures (a stack and a double-ended queue), in order to give the parser access to two focus items: the respective tops of the stack and the dequeue, that may or may not be adjacent. A dedicated action, GAP, pushes the top of the stack onto the bottom of the queue to make the next item in the stack available for a reduction.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "5"
            },
            {
                "text": "The split stack associated with the GAP action can be interpreted as a linear-access memory: it is possible to access the i th element in the stack, but it requires i operations.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "5"
            },
            {
                "text": "Non-local transitions Non-local transitions generalize standard parsing actions to nonadjacent elements in the parsing configurations. Maier and Lichte (2016) introduced a non-local transition SKIPSHIFT-i which applies SHIFT to the i th element in the buffer. Non-local transitions are also widely used in non-projective dependency parsing (Attardi, 2006; Qi and Manning, 2017; Fern\u00e1ndez-Gonz\u00e1lez and G\u00f3mez-Rodr\u00edguez, 2018 ).",
                "cite_spans": [
                    {
                        "start": 135,
                        "end": 158,
                        "text": "Maier and Lichte (2016)",
                        "ref_id": "BIBREF32"
                    },
                    {
                        "start": 340,
                        "end": 355,
                        "text": "(Attardi, 2006;",
                        "ref_id": "BIBREF0"
                    },
                    {
                        "start": 356,
                        "end": 377,
                        "text": "Qi and Manning, 2017;",
                        "ref_id": "BIBREF38"
                    },
                    {
                        "start": 378,
                        "end": 422,
                        "text": "Fern\u00e1ndez-Gonz\u00e1lez and G\u00f3mez-Rodr\u00edguez, 2018",
                        "ref_id": "BIBREF18"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "5"
            },
            {
                "text": "The key difference between these systems and ours is that we use an unordered memory. As a result, the semantics of the COMBINE-s action we introduce in Section 2 is independent from a specific position in the stack or the buffer. A system with an action such as SKIPSHIFT-i needs to learn parameters with every possible i, and will only learn parameters with the SKIPSHIFT-i actions that are required to derive the training set. In contrast, we use the same parameters to score each possible COMBINE-s action.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "5"
            },
            {
                "text": "We have presented a novel transition system that dispenses with the use of a stack, i.e. a memory with linear sequential access. Instead, the memory of the parser is represented by an unordered data structure with random-access: a set. We have designed a dynamic oracle for the resulting system and shown their empirical potential with state-of-the-art results on discontinuous constituency parsing of one English and two German treebanks. Finally, we plan to adapt our system to non-projective dependency parsing and semantic graph parsing. Table 7 : Detailed results. Overall, the positive effect of the dynamic oracle on F-score is explained by its effect on precision.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 542,
                        "end": 549,
                        "text": "Table 7",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Conclusion",
                "sec_num": "6"
            },
            {
                "text": "In such a case, we would need to score |S| \u00d7 |N | + 1 actions, where N is the set of nonterminals, instead of |S|+1 actions for our system.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "https://github.com/andreasvc/ disco-dop",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            }
        ],
        "back_matter": [
            {
                "text": "We thank Caio Corro, Giorgio Satta, Marco Damonte, as well as NAACL anonymous reviewers for feedback and suggestions. We gratefully acknowledge the support of Huawei Technologies.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Acknowledgments",
                "sec_num": null
            },
            {
                "text": "The oracle o leads to the reachable tree with the highest F-score with respect to the gold tree. The F-score of a predicted treet (represented as a set of instantiated constituents) with respect to a gold tree t * is defined as:By definition, o odd is optimal for precision because it constructs a constituent only if it is gold, and optimal for recall because it will construct a gold constituent if it is possible to do so. Moreover, o even is optimal for recall because any gold constituent reachable from c will still be reachable after any transition in o even (c, t * ). Assuming a configuration c = (S, s f , i, C) and next(c, t * ) = s g , we consider separately the SHIFT case and the COMBINE-s case: \u2022 COMBINE-s case: Let (X, s ) be a reachable gold constituent. Since it is compatible with s g , there are three possible cases: Finally, since o even does not construct new constituents (it is the role of labelling actions), it is optimal for precision.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Oracle Correctness",
                "sec_num": null
            },
            {
                "text": "The list of hyperparameters is presented in Table 5.\u2022 We use learning rate warm-up (linear increase from 0 to t 1000 during the first 1000 steps). \u2022 Before the t th update, we add Gaussian noise to the gradient of every parameter with mean 0 and variance 0.01 (1 + t) 0.55 (Neelakantan et al., 2015).\u2022 All experiments use greedy search decoding (we did not experiment with beam search). (Maier, 2015) Java, perceptron, beam=8, CPU 80 rparse (Maier, 2015) Java, perceptron, beam=1, CPU 640 Corro et al. (2017) C++, neural, CPU \u2248 7.3 Table 6 : Parsing times on development sets in tokens per second (tok/s) and sentences per second (sent/s). The parsing times are presented as reported by authors, they are not comparable across parsers (since the experiments were run on different hardware). Our parser is run on a single core of an Intel i7 CPU.\u2022 Before each training step, we replace a word embedding by an 'UNK' pseudo-word embedding with probability 0.3. We only do this replacement for the least frequent word-types ( 2 3 least frequent word-types). The 'UNK' embedding is then used to represent unknown words. \u2022 We apply dropout at the input of the tagger and the input of action scorers: each single prediction has its own dropout mask.",
                "cite_spans": [
                    {
                        "start": 387,
                        "end": 400,
                        "text": "(Maier, 2015)",
                        "ref_id": "BIBREF31"
                    },
                    {
                        "start": 441,
                        "end": 454,
                        "text": "(Maier, 2015)",
                        "ref_id": "BIBREF31"
                    },
                    {
                        "start": 489,
                        "end": 508,
                        "text": "Corro et al. (2017)",
                        "ref_id": "BIBREF10"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 532,
                        "end": 539,
                        "text": "Table 6",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "B Hyperparameters",
                "sec_num": null
            },
            {
                "text": "We report indicative parsing times in Table 6 . Detailed parsing results -including precision, recall, F-score and part-of-speech tag accuracy-are presented in Table 7 .",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 38,
                        "end": 45,
                        "text": "Table 6",
                        "ref_id": null
                    },
                    {
                        "start": 160,
                        "end": 167,
                        "text": "Table 7",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "C Detailed Results",
                "sec_num": null
            }
        ],
        "bib_entries": {
            "BIBREF0": {
                "ref_id": "b0",
                "title": "Experiments with a multilanguage non-projective dependency parser",
                "authors": [
                    {
                        "first": "Giuseppe",
                        "middle": [],
                        "last": "Attardi",
                        "suffix": ""
                    }
                ],
                "year": 2006,
                "venue": "Proceedings of the Tenth Conference on Computational Natural Language Learning (CoNLL-X)",
                "volume": "",
                "issue": "",
                "pages": "166--170",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Giuseppe Attardi. 2006. Experiments with a multilan- guage non-projective dependency parser. In Pro- ceedings of the Tenth Conference on Computational Natural Language Learning (CoNLL-X), pages 166- 170. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF1": {
                "ref_id": "b1",
                "title": "Training with exploration improves a greedy stack lstm parser",
                "authors": [
                    {
                        "first": "Miguel",
                        "middle": [],
                        "last": "Ballesteros",
                        "suffix": ""
                    },
                    {
                        "first": "Yoav",
                        "middle": [],
                        "last": "Goldberg",
                        "suffix": ""
                    },
                    {
                        "first": "Chris",
                        "middle": [],
                        "last": "Dyer",
                        "suffix": ""
                    },
                    {
                        "first": "Noah",
                        "middle": [
                            "A"
                        ],
                        "last": "Smith",
                        "suffix": ""
                    }
                ],
                "year": 2005,
                "venue": "Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {
                    "DOI": [
                        "10.18653/v1/D16-1211"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Miguel Ballesteros, Yoav Goldberg, Chris Dyer, and Noah A. Smith. 2016. Training with exploration im- proves a greedy stack lstm parser. In Proceedings of the 2016 Conference on Empirical Methods in Natu- ral Language Processing, pages 2005-2010, Austin, Texas. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF2": {
                "ref_id": "b2",
                "title": "Large-scale machine learning with stochastic gradient descent",
                "authors": [
                    {
                        "first": "L\u00e9on",
                        "middle": [],
                        "last": "Bottou",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "Proceedings of the 19th International Conference on Computational Statistics (COMPSTAT'2010)",
                "volume": "",
                "issue": "",
                "pages": "177--187",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "L\u00e9on Bottou. 2010. Large-scale machine learning with stochastic gradient descent. In Proceedings of the 19th International Conference on Compu- tational Statistics (COMPSTAT'2010), pages 177- 187, Paris, France. Springer.",
                "links": null
            },
            "BIBREF3": {
                "ref_id": "b3",
                "title": "Tiger: Linguistic interpretation of a german corpus",
                "authors": [
                    {
                        "first": "Sabine",
                        "middle": [],
                        "last": "Brants",
                        "suffix": ""
                    },
                    {
                        "first": "Stefanie",
                        "middle": [],
                        "last": "Dipper",
                        "suffix": ""
                    },
                    {
                        "first": "Peter",
                        "middle": [],
                        "last": "Eisenberg",
                        "suffix": ""
                    },
                    {
                        "first": "Silvia",
                        "middle": [],
                        "last": "Hansen-Schirra",
                        "suffix": ""
                    },
                    {
                        "first": "Esther",
                        "middle": [],
                        "last": "K\u00f6nig",
                        "suffix": ""
                    },
                    {
                        "first": "Wolfgang",
                        "middle": [],
                        "last": "Lezius",
                        "suffix": ""
                    },
                    {
                        "first": "Christian",
                        "middle": [],
                        "last": "Rohrer",
                        "suffix": ""
                    },
                    {
                        "first": "George",
                        "middle": [],
                        "last": "Smith",
                        "suffix": ""
                    },
                    {
                        "first": "Hans",
                        "middle": [],
                        "last": "Uszkoreit",
                        "suffix": ""
                    }
                ],
                "year": 2004,
                "venue": "Research on Language and Computation",
                "volume": "2",
                "issue": "4",
                "pages": "597--620",
                "other_ids": {
                    "DOI": [
                        "10.1007/s11168-004-7431-3"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Sabine Brants, Stefanie Dipper, Peter Eisenberg, Sil- via Hansen-Schirra, Esther K\u00f6nig, Wolfgang Lezius, Christian Rohrer, George Smith, and Hans Uszkor- eit. 2004. Tiger: Linguistic interpretation of a ger- man corpus. Research on Language and Computa- tion, 2(4):597-620.",
                "links": null
            },
            "BIBREF4": {
                "ref_id": "b4",
                "title": "Multitask learning",
                "authors": [
                    {
                        "first": "Rich",
                        "middle": [],
                        "last": "Caruana",
                        "suffix": ""
                    }
                ],
                "year": 1997,
                "venue": "Machine Learning",
                "volume": "28",
                "issue": "",
                "pages": "41--75",
                "other_ids": {
                    "DOI": [
                        "10.1023/A:1007379606734"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Rich Caruana. 1997. Multitask learning. Machine Learning, 28(1):41-75.",
                "links": null
            },
            "BIBREF5": {
                "ref_id": "b5",
                "title": "Neural greedy constituent parsing with dynamic oracles",
                "authors": [
                    {
                        "first": "Maximin",
                        "middle": [],
                        "last": "Coavoux",
                        "suffix": ""
                    },
                    {
                        "first": "Benoit",
                        "middle": [],
                        "last": "Crabb\u00e9",
                        "suffix": ""
                    }
                ],
                "year": 2016,
                "venue": "Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics",
                "volume": "1",
                "issue": "",
                "pages": "172--182",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Maximin Coavoux and Benoit Crabb\u00e9. 2016. Neural greedy constituent parsing with dynamic oracles. In Proceedings of the 54th Annual Meeting of the As- sociation for Computational Linguistics (Volume 1: Long Papers), pages 172-182, Berlin, Germany. As- sociation for Computational Linguistics.",
                "links": null
            },
            "BIBREF6": {
                "ref_id": "b6",
                "title": "Incremental discontinuous phrase structure parsing with the gap transition",
                "authors": [
                    {
                        "first": "Maximin",
                        "middle": [],
                        "last": "Coavoux",
                        "suffix": ""
                    },
                    {
                        "first": "Benoit",
                        "middle": [],
                        "last": "Crabb\u00e9",
                        "suffix": ""
                    }
                ],
                "year": 2017,
                "venue": "Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics",
                "volume": "1",
                "issue": "",
                "pages": "1259--1270",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Maximin Coavoux and Benoit Crabb\u00e9. 2017a. Incre- mental discontinuous phrase structure parsing with the gap transition. In Proceedings of the 15th Con- ference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Pa- pers, pages 1259-1270, Valencia, Spain. Associa- tion for Computational Linguistics.",
                "links": null
            },
            "BIBREF7": {
                "ref_id": "b7",
                "title": "Multilingual lexicalized constituency parsing with wordlevel auxiliary tasks",
                "authors": [
                    {
                        "first": "Maximin",
                        "middle": [],
                        "last": "Coavoux",
                        "suffix": ""
                    },
                    {
                        "first": "Benoit",
                        "middle": [],
                        "last": "Crabb\u00e9",
                        "suffix": ""
                    }
                ],
                "year": 2017,
                "venue": "Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics",
                "volume": "2",
                "issue": "",
                "pages": "331--336",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Maximin Coavoux and Benoit Crabb\u00e9. 2017b. Multi- lingual lexicalized constituency parsing with word- level auxiliary tasks. In Proceedings of the 15th Conference of the European Chapter of the Associa- tion for Computational Linguistics: Volume 2, Short Papers, pages 331-336, Valencia, Spain. Associa- tion for Computational Linguistics.",
                "links": null
            },
            "BIBREF8": {
                "ref_id": "b8",
                "title": "Unlexicalized transition-based discontinuous constituency parsing",
                "authors": [
                    {
                        "first": "Maximin",
                        "middle": [],
                        "last": "Coavoux",
                        "suffix": ""
                    },
                    {
                        "first": "Beno\u00eet",
                        "middle": [],
                        "last": "Crabb\u00e9",
                        "suffix": ""
                    },
                    {
                        "first": "Shay",
                        "middle": [
                            "B"
                        ],
                        "last": "Cohen",
                        "suffix": ""
                    }
                ],
                "year": 2019,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Maximin Coavoux, Beno\u00eet Crabb\u00e9, and Shay B. Cohen. 2019. Unlexicalized transition-based discontinuous constituency parsing. CoRR, abs/1902.08912v1.",
                "links": null
            },
            "BIBREF9": {
                "ref_id": "b9",
                "title": "Elimination of spurious ambiguity in transition-based dependency parsing",
                "authors": [
                    {
                        "first": "B",
                        "middle": [],
                        "last": "Shay",
                        "suffix": ""
                    },
                    {
                        "first": "Carlos",
                        "middle": [],
                        "last": "Cohen",
                        "suffix": ""
                    },
                    {
                        "first": "Giorgio",
                        "middle": [],
                        "last": "G\u00f3mez-Rodr\u00edguez",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Satta",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Shay B. Cohen, Carlos G\u00f3mez-Rodr\u00edguez, and Gior- gio Satta. 2012. Elimination of spurious ambigu- ity in transition-based dependency parsing. CoRR, abs/1206.6735.",
                "links": null
            },
            "BIBREF10": {
                "ref_id": "b10",
                "title": "Efficient discontinuous phrase-structure parsing via the generalized maximum spanning arborescence",
                "authors": [
                    {
                        "first": "Caio",
                        "middle": [],
                        "last": "Corro",
                        "suffix": ""
                    },
                    {
                        "first": "Joseph",
                        "middle": [
                            "Le"
                        ],
                        "last": "Roux",
                        "suffix": ""
                    },
                    {
                        "first": "Mathieu",
                        "middle": [],
                        "last": "Lacroix",
                        "suffix": ""
                    }
                ],
                "year": 2017,
                "venue": "Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "1645--1655",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Caio Corro, Joseph Le Roux, and Mathieu Lacroix. 2017. Efficient discontinuous phrase-structure pars- ing via the generalized maximum spanning arbores- cence. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Process- ing, pages 1645-1655, Copenhagen, Denmark. As- sociation for Computational Linguistics.",
                "links": null
            },
            "BIBREF11": {
                "ref_id": "b11",
                "title": "Multilingual discriminative lexicalized phrase structure parsing",
                "authors": [
                    {
                        "first": "",
                        "middle": [],
                        "last": "Benoit Crabb\u00e9",
                        "suffix": ""
                    }
                ],
                "year": 2015,
                "venue": "Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "1847--1856",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Benoit Crabb\u00e9. 2015. Multilingual discriminative lex- icalized phrase structure parsing. In Proceedings of the 2015 Conference on Empirical Methods in Nat- ural Language Processing, pages 1847-1856, Lis- bon, Portugal. Association for Computational Lin- guistics.",
                "links": null
            },
            "BIBREF12": {
                "ref_id": "b12",
                "title": "Data-oriented parsing with discontinuous constituents and function tags",
                "authors": [
                    {
                        "first": "Andreas",
                        "middle": [],
                        "last": "Van Cranenburgh",
                        "suffix": ""
                    },
                    {
                        "first": "Remko",
                        "middle": [],
                        "last": "Scha",
                        "suffix": ""
                    },
                    {
                        "first": "Rens",
                        "middle": [],
                        "last": "Bod",
                        "suffix": ""
                    }
                ],
                "year": 2016,
                "venue": "Journal of Language Modelling",
                "volume": "4",
                "issue": "1",
                "pages": "57--111",
                "other_ids": {
                    "DOI": [
                        "10.15398/jlm.v4i1.100"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Andreas van Cranenburgh, Remko Scha, and Rens Bod. 2016. Data-oriented parsing with discontinu- ous constituents and function tags. Journal of Lan- guage Modelling, 4(1):57-111.",
                "links": null
            },
            "BIBREF13": {
                "ref_id": "b13",
                "title": "Incremental parsing with minimal features using bi-directional LSTM",
                "authors": [
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Cross",
                        "suffix": ""
                    },
                    {
                        "first": "Liang",
                        "middle": [],
                        "last": "Huang",
                        "suffix": ""
                    }
                ],
                "year": 2016,
                "venue": "Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics",
                "volume": "2",
                "issue": "",
                "pages": "32--37",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "James Cross and Liang Huang. 2016a. Incremental parsing with minimal features using bi-directional LSTM. In Proceedings of the 54th Annual Meet- ing of the Association for Computational Linguistics (Volume 2: Short Papers), pages 32-37, Berlin, Ger- many. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF14": {
                "ref_id": "b14",
                "title": "Span-based constituency parsing with a structure-label system and provably optimal dynamic oracles",
                "authors": [
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Cross",
                        "suffix": ""
                    },
                    {
                        "first": "Liang",
                        "middle": [],
                        "last": "Huang",
                        "suffix": ""
                    }
                ],
                "year": 2016,
                "venue": "Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "1--11",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "James Cross and Liang Huang. 2016b. Span-based constituency parsing with a structure-label system and provably optimal dynamic oracles. In Proceed- ings of the 2016 Conference on Empirical Meth- ods in Natural Language Processing, pages 1-11, Austin, Texas. Association for Computational Lin- guistics.",
                "links": null
            },
            "BIBREF15": {
                "ref_id": "b15",
                "title": "Probabilistic parsing for german using sister-head dependencies",
                "authors": [
                    {
                        "first": "Amit",
                        "middle": [],
                        "last": "Dubey",
                        "suffix": ""
                    },
                    {
                        "first": "Frank",
                        "middle": [],
                        "last": "Keller",
                        "suffix": ""
                    }
                ],
                "year": 2003,
                "venue": "Proceedings of the 41st Annual Meeting of the Association for Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "96--103",
                "other_ids": {
                    "DOI": [
                        "10.3115/1075096.1075109"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Amit Dubey and Frank Keller. 2003. Probabilistic parsing for german using sister-head dependencies. In Proceedings of the 41st Annual Meeting of the As- sociation for Computational Linguistics, pages 96- 103, Sapporo, Japan. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF16": {
                "ref_id": "b16",
                "title": "Neural CRF parsing",
                "authors": [
                    {
                        "first": "Greg",
                        "middle": [],
                        "last": "Durrett",
                        "suffix": ""
                    },
                    {
                        "first": "Dan",
                        "middle": [],
                        "last": "Klein",
                        "suffix": ""
                    }
                ],
                "year": 2015,
                "venue": "Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing",
                "volume": "1",
                "issue": "",
                "pages": "302--312",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Greg Durrett and Dan Klein. 2015. Neural CRF pars- ing. In Proceedings of the 53rd Annual Meet- ing of the Association for Computational Linguis- tics and the 7th International Joint Conference on Natural Language Processing (Volume 1: Long Pa- pers), pages 302-312, Beijing, China. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF17": {
                "ref_id": "b17",
                "title": "PLCFRS parsing of english discontinuous constituents",
                "authors": [
                    {
                        "first": "Kilian",
                        "middle": [],
                        "last": "Evang",
                        "suffix": ""
                    },
                    {
                        "first": "Laura",
                        "middle": [],
                        "last": "Kallmeyer",
                        "suffix": ""
                    }
                ],
                "year": 2011,
                "venue": "Proceedings of the 12th International Conference on Parsing Technologies",
                "volume": "",
                "issue": "",
                "pages": "104--116",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Kilian Evang and Laura Kallmeyer. 2011. PLCFRS parsing of english discontinuous constituents. In Proceedings of the 12th International Conference on Parsing Technologies, pages 104-116, Dublin, Ire- land. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF18": {
                "ref_id": "b18",
                "title": "Non-projective dependency parsing with non-local transitions",
                "authors": [
                    {
                        "first": "Daniel",
                        "middle": [],
                        "last": "Fern\u00e1ndez",
                        "suffix": ""
                    },
                    {
                        "first": "-",
                        "middle": [],
                        "last": "Gonz\u00e1lez",
                        "suffix": ""
                    },
                    {
                        "first": "Carlos",
                        "middle": [],
                        "last": "G\u00f3mez-Rodr\u00edguez",
                        "suffix": ""
                    }
                ],
                "year": 2018,
                "venue": "Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",
                "volume": "2",
                "issue": "",
                "pages": "693--700",
                "other_ids": {
                    "DOI": [
                        "10.18653/v1/N18-2109"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Daniel Fern\u00e1ndez-Gonz\u00e1lez and Carlos G\u00f3mez- Rodr\u00edguez. 2018. Non-projective dependency parsing with non-local transitions. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Lin- guistics: Human Language Technologies, Volume 2 (Short Papers), pages 693-700. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF19": {
                "ref_id": "b19",
                "title": "Dynamic oracles for top-down and in-order shift-reduce constituent parsing",
                "authors": [
                    {
                        "first": "Daniel",
                        "middle": [],
                        "last": "Fern\u00e1ndez",
                        "suffix": ""
                    },
                    {
                        "first": "-",
                        "middle": [],
                        "last": "Gonz\u00e1lez",
                        "suffix": ""
                    },
                    {
                        "first": "Carlos",
                        "middle": [],
                        "last": "G\u00f3mez-Rodr\u00edguez",
                        "suffix": ""
                    }
                ],
                "year": 2018,
                "venue": "Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "1303--1313",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Daniel Fern\u00e1ndez-Gonz\u00e1lez and Carlos G\u00f3mez- Rodr\u00edguez. 2018a. Dynamic oracles for top-down and in-order shift-reduce constituent parsing. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 1303-1313, Brussels, Belgium. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF20": {
                "ref_id": "b20",
                "title": "Faster shift-reduce constituent parsing with a non-binary, bottom-up strategy",
                "authors": [
                    {
                        "first": "Daniel",
                        "middle": [],
                        "last": "Fern\u00e1ndez",
                        "suffix": ""
                    },
                    {
                        "first": "-",
                        "middle": [],
                        "last": "Gonz\u00e1lez",
                        "suffix": ""
                    },
                    {
                        "first": "Carlos",
                        "middle": [],
                        "last": "G\u00f3mez-Rodr\u00edguez",
                        "suffix": ""
                    }
                ],
                "year": 2018,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Daniel Fern\u00e1ndez-Gonz\u00e1lez and Carlos G\u00f3mez- Rodr\u00edguez. 2018b. Faster shift-reduce constituent parsing with a non-binary, bottom-up strategy. CoRR, abs/1804.07961.",
                "links": null
            },
            "BIBREF21": {
                "ref_id": "b21",
                "title": "Parsing as reduction",
                "authors": [
                    {
                        "first": "Daniel",
                        "middle": [],
                        "last": "Fern\u00e1ndez",
                        "suffix": ""
                    },
                    {
                        "first": "-Gonz\u00e1lez",
                        "middle": [],
                        "last": "Andr\u00e9",
                        "suffix": ""
                    },
                    {
                        "first": "F",
                        "middle": [
                            "T"
                        ],
                        "last": "Martins",
                        "suffix": ""
                    }
                ],
                "year": 2015,
                "venue": "Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "1523--1533",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Daniel Fern\u00e1ndez-Gonz\u00e1lez and Andr\u00e9 F. T. Martins. 2015. Parsing as reduction. In Proceedings of the 53rd Annual Meeting of the Association for Compu- tational Linguistics and the 7th International Joint Conference on Natural Language Processing (Vol- ume 1: Long Papers), pages 1523-1533, Beijing, China. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF22": {
                "ref_id": "b22",
                "title": "Generic refinement of expressive grammar formalisms with an application to discontinuous constituent parsing",
                "authors": [
                    {
                        "first": "Kilian",
                        "middle": [],
                        "last": "Gebhardt",
                        "suffix": ""
                    }
                ],
                "year": 2018,
                "venue": "Proceedings of the 27th International Conference on Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "3049--3063",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Kilian Gebhardt. 2018. Generic refinement of expres- sive grammar formalisms with an application to dis- continuous constituent parsing. In Proceedings of the 27th International Conference on Computational Linguistics, pages 3049-3063, Santa Fe, New Mex- ico, USA. Association for Computational Linguis- tics.",
                "links": null
            },
            "BIBREF23": {
                "ref_id": "b23",
                "title": "Understanding the difficulty of training deep feedforward neural networks",
                "authors": [
                    {
                        "first": "Xavier",
                        "middle": [],
                        "last": "Glorot",
                        "suffix": ""
                    },
                    {
                        "first": "Yoshua",
                        "middle": [],
                        "last": "Bengio",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "Proceedings of the International Conference on Artificial Intelligence and Statistics (AISTATS10). Society for Artificial Intelligence and Statistics",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Xavier Glorot and Yoshua Bengio. 2010. Understand- ing the difficulty of training deep feedforward neu- ral networks. In In Proceedings of the International Conference on Artificial Intelligence and Statistics (AISTATS10). Society for Artificial Intelligence and Statistics.",
                "links": null
            },
            "BIBREF24": {
                "ref_id": "b24",
                "title": "A dynamic oracle for arc-eager dependency parsing",
                "authors": [
                    {
                        "first": "Yoav",
                        "middle": [],
                        "last": "Goldberg",
                        "suffix": ""
                    },
                    {
                        "first": "Joakim",
                        "middle": [],
                        "last": "Nivre",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "The COLING 2012 Organizing Committee",
                "volume": "",
                "issue": "",
                "pages": "959--976",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Yoav Goldberg and Joakim Nivre. 2012. A dynamic or- acle for arc-eager dependency parsing. In Proceed- ings of COLING 2012, pages 959-976, Mumbai, In- dia. The COLING 2012 Organizing Committee.",
                "links": null
            },
            "BIBREF25": {
                "ref_id": "b25",
                "title": "Training deterministic parsers with non-deterministic oracles",
                "authors": [
                    {
                        "first": "Yoav",
                        "middle": [],
                        "last": "Goldberg",
                        "suffix": ""
                    },
                    {
                        "first": "Joakim",
                        "middle": [],
                        "last": "Nivre",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "Transactions of the Association for Computational Linguistics",
                "volume": "1",
                "issue": "",
                "pages": "403--414",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Yoav Goldberg and Joakim Nivre. 2013. Training de- terministic parsers with non-deterministic oracles. Transactions of the Association for Computational Linguistics, 1:403-414.",
                "links": null
            },
            "BIBREF26": {
                "ref_id": "b26",
                "title": "An efficient dynamic oracle for unrestricted non-projective parsing",
                "authors": [
                    {
                        "first": "Carlos",
                        "middle": [],
                        "last": "G\u00f3mez",
                        "suffix": ""
                    },
                    {
                        "first": "-Rodr\u00edguez",
                        "middle": [],
                        "last": "",
                        "suffix": ""
                    },
                    {
                        "first": "Daniel",
                        "middle": [],
                        "last": "Fern\u00e1ndez-Gonz\u00e1lez",
                        "suffix": ""
                    }
                ],
                "year": 2015,
                "venue": "Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing",
                "volume": "2",
                "issue": "",
                "pages": "256--261",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Carlos G\u00f3mez-Rodr\u00edguez and Daniel Fern\u00e1ndez- Gonz\u00e1lez. 2015. An efficient dynamic oracle for unrestricted non-projective parsing. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 2: Short Papers), pages 256-261, Beijing, China. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF27": {
                "ref_id": "b27",
                "title": "A polynomial-time dynamic oracle for non-projective dependency parsing",
                "authors": [
                    {
                        "first": "Carlos",
                        "middle": [],
                        "last": "G\u00f3mez-Rodr\u00edguez",
                        "suffix": ""
                    },
                    {
                        "first": "Francesco",
                        "middle": [],
                        "last": "Sartorio",
                        "suffix": ""
                    },
                    {
                        "first": "Giorgio",
                        "middle": [],
                        "last": "Satta",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
                "volume": "",
                "issue": "",
                "pages": "917--927",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Carlos G\u00f3mez-Rodr\u00edguez, Francesco Sartorio, and Giorgio Satta. 2014. A polynomial-time dy- namic oracle for non-projective dependency pars- ing. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 917-927, Doha, Qatar. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF28": {
                "ref_id": "b28",
                "title": "Less grammar, more features",
                "authors": [
                    {
                        "first": "David",
                        "middle": [],
                        "last": "Hall",
                        "suffix": ""
                    },
                    {
                        "first": "Greg",
                        "middle": [],
                        "last": "Durrett",
                        "suffix": ""
                    },
                    {
                        "first": "Dan",
                        "middle": [],
                        "last": "Klein",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics",
                "volume": "1",
                "issue": "",
                "pages": "228--237",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "David Hall, Greg Durrett, and Dan Klein. 2014. Less grammar, more features. In Proceedings of the 52nd Annual Meeting of the Association for Com- putational Linguistics (Volume 1: Long Papers), pages 228-237, Baltimore, Maryland. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF29": {
                "ref_id": "b29",
                "title": "Parsing Beyond Context-Free Grammars",
                "authors": [
                    {
                        "first": "Laura",
                        "middle": [],
                        "last": "Kallmeyer",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Laura Kallmeyer. 2010. Parsing Beyond Context-Free Grammars, 1st edition. Springer Publishing Com- pany, Incorporated.",
                "links": null
            },
            "BIBREF30": {
                "ref_id": "b30",
                "title": "Simple and accurate dependency parsing using bidirectional LSTM feature representations",
                "authors": [
                    {
                        "first": "Eliyahu",
                        "middle": [],
                        "last": "Kiperwasser",
                        "suffix": ""
                    },
                    {
                        "first": "Yoav",
                        "middle": [],
                        "last": "Goldberg",
                        "suffix": ""
                    }
                ],
                "year": 2016,
                "venue": "Transactions of the Association for Computational Linguistics",
                "volume": "4",
                "issue": "",
                "pages": "313--327",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Eliyahu Kiperwasser and Yoav Goldberg. 2016. Sim- ple and accurate dependency parsing using bidirec- tional LSTM feature representations. Transactions of the Association for Computational Linguistics, 4:313-327.",
                "links": null
            },
            "BIBREF31": {
                "ref_id": "b31",
                "title": "Discontinuous incremental shift-reduce parsing",
                "authors": [
                    {
                        "first": "Wolfgang",
                        "middle": [],
                        "last": "Maier",
                        "suffix": ""
                    }
                ],
                "year": 2015,
                "venue": "Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing",
                "volume": "1",
                "issue": "",
                "pages": "1202--1212",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Wolfgang Maier. 2015. Discontinuous incremental shift-reduce parsing. In Proceedings of the 53rd An- nual Meeting of the Association for Computational Linguistics and the 7th International Joint Confer- ence on Natural Language Processing (Volume 1: Long Papers), pages 1202-1212, Beijing, China. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF32": {
                "ref_id": "b32",
                "title": "Discontinuous parsing with continuous trees",
                "authors": [
                    {
                        "first": "Wolfgang",
                        "middle": [],
                        "last": "Maier",
                        "suffix": ""
                    },
                    {
                        "first": "Timm",
                        "middle": [],
                        "last": "Lichte",
                        "suffix": ""
                    }
                ],
                "year": 2016,
                "venue": "Proceedings of the Workshop on Discontinuous Structures in Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "47--57",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Wolfgang Maier and Timm Lichte. 2016. Discontinu- ous parsing with continuous trees. In Proceedings of the Workshop on Discontinuous Structures in Natu- ral Language Processing, pages 47-57, San Diego, California. Association for Computational Linguis- tics.",
                "links": null
            },
            "BIBREF33": {
                "ref_id": "b33",
                "title": "Building a large annotated corpus of english: The penn treebank",
                "authors": [
                    {
                        "first": "Mitchell",
                        "middle": [],
                        "last": "Marcus",
                        "suffix": ""
                    },
                    {
                        "first": "Beatrice",
                        "middle": [],
                        "last": "Santorini",
                        "suffix": ""
                    },
                    {
                        "first": "Mary",
                        "middle": [
                            "Ann"
                        ],
                        "last": "Marcinkiewicz",
                        "suffix": ""
                    }
                ],
                "year": 1993,
                "venue": "Computational Linguistics",
                "volume": "19",
                "issue": "2",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Mitchell Marcus, Beatrice Santorini, and Mary Ann Marcinkiewicz. 1993. Building a large anno- tated corpus of english: The penn treebank. Com- putational Linguistics, Volume 19, Number 2, June 1993, Special Issue on Using Large Corpora: II.",
                "links": null
            },
            "BIBREF34": {
                "ref_id": "b34",
                "title": "Adding gradient noise improves learning for very deep networks",
                "authors": [
                    {
                        "first": "Arvind",
                        "middle": [],
                        "last": "Neelakantan",
                        "suffix": ""
                    },
                    {
                        "first": "Luke",
                        "middle": [],
                        "last": "Vilnis",
                        "suffix": ""
                    },
                    {
                        "first": "Quoc",
                        "middle": [
                            "V"
                        ],
                        "last": "Le",
                        "suffix": ""
                    },
                    {
                        "first": "Ilya",
                        "middle": [],
                        "last": "Sutskever",
                        "suffix": ""
                    },
                    {
                        "first": "Lukasz",
                        "middle": [],
                        "last": "Kaiser",
                        "suffix": ""
                    },
                    {
                        "first": "Karol",
                        "middle": [],
                        "last": "Kurach",
                        "suffix": ""
                    },
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Martens",
                        "suffix": ""
                    }
                ],
                "year": 2015,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Arvind Neelakantan, Luke Vilnis, Quoc V. Le, Ilya Sutskever, Lukasz Kaiser, Karol Kurach, and James Martens. 2015. Adding gradient noise im- proves learning for very deep networks. CoRR, abs/1511.06807.",
                "links": null
            },
            "BIBREF35": {
                "ref_id": "b35",
                "title": "Non-projective dependency parsing in expected linear time",
                "authors": [
                    {
                        "first": "Joakim",
                        "middle": [],
                        "last": "Nivre",
                        "suffix": ""
                    }
                ],
                "year": 2009,
                "venue": "Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP",
                "volume": "",
                "issue": "",
                "pages": "351--359",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Joakim Nivre. 2009. Non-projective dependency pars- ing in expected linear time. In Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP, pages 351-359, Suntec, Singapore. Association for Com- putational Linguistics.",
                "links": null
            },
            "BIBREF36": {
                "ref_id": "b36",
                "title": "Automatic differentiation in pytorch",
                "authors": [
                    {
                        "first": "Adam",
                        "middle": [],
                        "last": "Paszke",
                        "suffix": ""
                    },
                    {
                        "first": "Sam",
                        "middle": [],
                        "last": "Gross",
                        "suffix": ""
                    },
                    {
                        "first": "Soumith",
                        "middle": [],
                        "last": "Chintala",
                        "suffix": ""
                    },
                    {
                        "first": "Gregory",
                        "middle": [],
                        "last": "Chanan",
                        "suffix": ""
                    },
                    {
                        "first": "Edward",
                        "middle": [],
                        "last": "Yang",
                        "suffix": ""
                    },
                    {
                        "first": "Zachary",
                        "middle": [],
                        "last": "Devito",
                        "suffix": ""
                    },
                    {
                        "first": "Zeming",
                        "middle": [],
                        "last": "Lin",
                        "suffix": ""
                    },
                    {
                        "first": "Alban",
                        "middle": [],
                        "last": "Desmaison",
                        "suffix": ""
                    },
                    {
                        "first": "Luca",
                        "middle": [],
                        "last": "Antiga",
                        "suffix": ""
                    },
                    {
                        "first": "Adam",
                        "middle": [],
                        "last": "Lerer",
                        "suffix": ""
                    }
                ],
                "year": 2017,
                "venue": "NIPS-W",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Adam Paszke, Sam Gross, Soumith Chintala, Gre- gory Chanan, Edward Yang, Zachary DeVito, Zem- ing Lin, Alban Desmaison, Luca Antiga, and Adam Lerer. 2017. Automatic differentiation in pytorch. In NIPS-W.",
                "links": null
            },
            "BIBREF37": {
                "ref_id": "b37",
                "title": "Acceleration of stochastic approximation by averaging",
                "authors": [
                    {
                        "first": "T",
                        "middle": [],
                        "last": "Boris",
                        "suffix": ""
                    },
                    {
                        "first": "Anatoli",
                        "middle": [
                            "B"
                        ],
                        "last": "Polyak",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Juditsky",
                        "suffix": ""
                    }
                ],
                "year": 1992,
                "venue": "SIAM Journal on Control and Optimization",
                "volume": "30",
                "issue": "4",
                "pages": "838--855",
                "other_ids": {
                    "DOI": [
                        "10.1137/0330046"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Boris T. Polyak and Anatoli B. Juditsky. 1992. Ac- celeration of stochastic approximation by averag- ing. SIAM Journal on Control and Optimization, 30(4):838-855.",
                "links": null
            },
            "BIBREF38": {
                "ref_id": "b38",
                "title": "Arc-swift: A novel transition system for dependency parsing",
                "authors": [
                    {
                        "first": "Peng",
                        "middle": [],
                        "last": "Qi",
                        "suffix": ""
                    },
                    {
                        "first": "D",
                        "middle": [],
                        "last": "Christopher",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Manning",
                        "suffix": ""
                    }
                ],
                "year": 2017,
                "venue": "Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics",
                "volume": "2",
                "issue": "",
                "pages": "110--117",
                "other_ids": {
                    "DOI": [
                        "10.18653/v1/P17-2018"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Peng Qi and Christopher D. Manning. 2017. Arc-swift: A novel transition system for dependency parsing. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 110-117. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF39": {
                "ref_id": "b39",
                "title": "Overview of the SPMRL 2013 shared task: A cross-framework evaluation of parsing morphologically rich languages",
                "authors": [
                    {
                        "first": "Djam\u00e9",
                        "middle": [],
                        "last": "Seddah",
                        "suffix": ""
                    },
                    {
                        "first": "Reut",
                        "middle": [],
                        "last": "Tsarfaty",
                        "suffix": ""
                    },
                    {
                        "first": "Sandra",
                        "middle": [],
                        "last": "K\u00fcbler",
                        "suffix": ""
                    },
                    {
                        "first": "Marie",
                        "middle": [],
                        "last": "Candito",
                        "suffix": ""
                    },
                    {
                        "first": "Jinho",
                        "middle": [
                            "D"
                        ],
                        "last": "Choi",
                        "suffix": ""
                    },
                    {
                        "first": "Rich\u00e1rd",
                        "middle": [],
                        "last": "Farkas",
                        "suffix": ""
                    },
                    {
                        "first": "Jennifer",
                        "middle": [],
                        "last": "Foster",
                        "suffix": ""
                    },
                    {
                        "first": "Iakes",
                        "middle": [],
                        "last": "Goenaga",
                        "suffix": ""
                    },
                    {
                        "first": "Yoav",
                        "middle": [],
                        "last": "Koldo Gojenola Galletebeitia",
                        "suffix": ""
                    },
                    {
                        "first": "Spence",
                        "middle": [],
                        "last": "Goldberg",
                        "suffix": ""
                    },
                    {
                        "first": "Nizar",
                        "middle": [],
                        "last": "Green",
                        "suffix": ""
                    },
                    {
                        "first": "Marco",
                        "middle": [],
                        "last": "Habash",
                        "suffix": ""
                    },
                    {
                        "first": "Wolfgang",
                        "middle": [],
                        "last": "Kuhlmann",
                        "suffix": ""
                    },
                    {
                        "first": "Joakim",
                        "middle": [],
                        "last": "Maier",
                        "suffix": ""
                    },
                    {
                        "first": "Adam",
                        "middle": [],
                        "last": "Nivre",
                        "suffix": ""
                    },
                    {
                        "first": "Ryan",
                        "middle": [],
                        "last": "Przepi\u00f3rkowski",
                        "suffix": ""
                    },
                    {
                        "first": "Wolfgang",
                        "middle": [],
                        "last": "Roth",
                        "suffix": ""
                    },
                    {
                        "first": "Yannick",
                        "middle": [],
                        "last": "Seeker",
                        "suffix": ""
                    },
                    {
                        "first": "Veronika",
                        "middle": [],
                        "last": "Versley",
                        "suffix": ""
                    },
                    {
                        "first": "Marcin",
                        "middle": [],
                        "last": "Vincze",
                        "suffix": ""
                    },
                    {
                        "first": "Alina",
                        "middle": [],
                        "last": "Woli\u0144ski",
                        "suffix": ""
                    },
                    {
                        "first": "Eric",
                        "middle": [],
                        "last": "Wr\u00f3blewska",
                        "suffix": ""
                    },
                    {
                        "first": "Clergerie",
                        "middle": [],
                        "last": "Villemonte De La",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "Proceedings of the Fourth Workshop on Statistical Parsing of Morphologically-Rich Languages",
                "volume": "",
                "issue": "",
                "pages": "146--182",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Djam\u00e9 Seddah, Reut Tsarfaty, Sandra K\u00fcbler, Marie Candito, Jinho D. Choi, Rich\u00e1rd Farkas, Jen- nifer Foster, Iakes Goenaga, Koldo Gojenola Gal- letebeitia, Yoav Goldberg, Spence Green, Nizar Habash, Marco Kuhlmann, Wolfgang Maier, Joakim Nivre, Adam Przepi\u00f3rkowski, Ryan Roth, Wolfgang Seeker, Yannick Versley, Veronika Vincze, Marcin Woli\u0144ski, Alina Wr\u00f3blewska, and Eric Villemonte de la Clergerie. 2013. Overview of the SPMRL 2013 shared task: A cross-framework evaluation of pars- ing morphologically rich languages. In Proceed- ings of the Fourth Workshop on Statistical Parsing of Morphologically-Rich Languages, pages 146-182, Seattle, Washington, USA. Association for Compu- tational Linguistics.",
                "links": null
            },
            "BIBREF40": {
                "ref_id": "b40",
                "title": "An annotation scheme for free word order languages",
                "authors": [
                    {
                        "first": "Wojciech",
                        "middle": [],
                        "last": "Skut",
                        "suffix": ""
                    },
                    {
                        "first": "Brigitte",
                        "middle": [],
                        "last": "Krenn",
                        "suffix": ""
                    },
                    {
                        "first": "Thorsten",
                        "middle": [],
                        "last": "Brants",
                        "suffix": ""
                    },
                    {
                        "first": "Hans",
                        "middle": [],
                        "last": "Uszkoreit",
                        "suffix": ""
                    }
                ],
                "year": 1997,
                "venue": "Proceedings of the Fifth Conference on Applied Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "88--95",
                "other_ids": {
                    "DOI": [
                        "10.3115/974557.974571"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Wojciech Skut, Brigitte Krenn, Thorsten Brants, and Hans Uszkoreit. 1997. An annotation scheme for free word order languages. In Proceedings of the Fifth Conference on Applied Natural Language Pro- cessing, pages 88-95, Washington, DC, USA. Asso- ciation for Computational Linguistics.",
                "links": null
            },
            "BIBREF41": {
                "ref_id": "b41",
                "title": "Neural discontinuous constituency parsing",
                "authors": [
                    {
                        "first": "Milo\u0161",
                        "middle": [],
                        "last": "Stanojevi\u0107",
                        "suffix": ""
                    },
                    {
                        "first": "Raquel",
                        "middle": [],
                        "last": "Garrido Alhama",
                        "suffix": ""
                    }
                ],
                "year": 2017,
                "venue": "Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "1667--1677",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Milo\u0161 Stanojevi\u0107 and Raquel Garrido Alhama. 2017. Neural discontinuous constituency parsing. In Pro- ceedings of the 2017 Conference on Empirical Meth- ods in Natural Language Processing, pages 1667- 1677, Copenhagen, Denmark. Association for Com- putational Linguistics.",
                "links": null
            },
            "BIBREF42": {
                "ref_id": "b42",
                "title": "A minimal span-based neural constituency parser",
                "authors": [
                    {
                        "first": "Mitchell",
                        "middle": [],
                        "last": "Stern",
                        "suffix": ""
                    },
                    {
                        "first": "Jacob",
                        "middle": [],
                        "last": "Andreas",
                        "suffix": ""
                    },
                    {
                        "first": "Dan",
                        "middle": [],
                        "last": "Klein",
                        "suffix": ""
                    }
                ],
                "year": 2017,
                "venue": "Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics",
                "volume": "1",
                "issue": "",
                "pages": "818--827",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Mitchell Stern, Jacob Andreas, and Dan Klein. 2017. A minimal span-based neural constituency parser. In Proceedings of the 55th Annual Meeting of the As- sociation for Computational Linguistics (Volume 1: Long Papers), pages 818-827, Vancouver, Canada. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF43": {
                "ref_id": "b43",
                "title": "Experiments with easy-first nonprojective constituent parsing",
                "authors": [
                    {
                        "first": "Yannick",
                        "middle": [],
                        "last": "Versley",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "Proceedings of the First Joint Workshop on Statistical Parsing of Morphologically Rich Languages and Syntactic Analysis of Non-Canonical Languages",
                "volume": "",
                "issue": "",
                "pages": "39--53",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Yannick Versley. 2014a. Experiments with easy-first nonprojective constituent parsing. In Proceedings of the First Joint Workshop on Statistical Parsing of Morphologically Rich Languages and Syntactic Analysis of Non-Canonical Languages, pages 39- 53, Dublin, Ireland. Dublin City University.",
                "links": null
            },
            "BIBREF44": {
                "ref_id": "b44",
                "title": "Incorporating semisupervised features into discontinuous easy-first constituent parsing",
                "authors": [
                    {
                        "first": "Yannick",
                        "middle": [],
                        "last": "Versley",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Yannick Versley. 2014b. Incorporating semi- supervised features into discontinuous easy-first constituent parsing. CoRR, abs/1409.3813.",
                "links": null
            },
            "BIBREF45": {
                "ref_id": "b45",
                "title": "Discontinuity re\u02c62-visited: A minimalist approach to pseudoprojective constituent parsing",
                "authors": [
                    {
                        "first": "Yannick",
                        "middle": [],
                        "last": "Versley",
                        "suffix": ""
                    }
                ],
                "year": 2016,
                "venue": "Proceedings of the Workshop on Discontinuous Structures in Natural Language Processing",
                "volume": "",
                "issue": "",
                "pages": "58--69",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Yannick Versley. 2016. Discontinuity re\u02c62-visited: A minimalist approach to pseudoprojective constituent parsing. In Proceedings of the Workshop on Discon- tinuous Structures in Natural Language Processing, pages 58-69, San Diego, California. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF46": {
                "ref_id": "b46",
                "title": "Characterizing structural descriptions produced by various grammatical formalisms",
                "authors": [
                    {
                        "first": "K",
                        "middle": [],
                        "last": "Vijay-Shanker",
                        "suffix": ""
                    },
                    {
                        "first": "David",
                        "middle": [
                            "J"
                        ],
                        "last": "Weir",
                        "suffix": ""
                    },
                    {
                        "first": "Aravind",
                        "middle": [
                            "K"
                        ],
                        "last": "Joshi",
                        "suffix": ""
                    }
                ],
                "year": 1987,
                "venue": "25th Annual Meeting of the Association for Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "K. Vijay-Shanker, David J. Weir, and Aravind K. Joshi. 1987. Characterizing structural descriptions pro- duced by various grammatical formalisms. In 25th Annual Meeting of the Association for Computa- tional Linguistics.",
                "links": null
            }
        },
        "ref_entries": {
            "FIGREF0": {
                "text": "Discontinuous constituency tree from the Discontinuous Penn treebank.",
                "num": null,
                "uris": null,
                "type_str": "figure"
            },
            "FIGREF1": {
                "text": "Initial configuration (\u2205, null, 0, \u2205) : 0 Goal configuration(\u2205, {0, 1, . . . , n \u2212 1}, n, C) : 4n \u2212 2 s f , i, C) : j \u21d2 (S \u222a {s f }, {i}, i + 1, C) : j + 1 i < n, j is even COMBINE-s (S, s f , i, C) : j \u21d2 (S \u2212 s, s f \u222a s, i, C) : j + 1 s \u2208 S, j is even Labelling actions LABEL-X (S, s f , i, C) : j \u21d2 (S, s f , i, C \u222a {(X, s f )}) : j + 1 j is odd NO-LABEL (S, s f , i, C) : j \u21d2 (S, s f , i, C) : j + 1 i = n or S = \u2205, j is odd",
                "num": null,
                "uris": null,
                "type_str": "figure"
            },
            "TABREF0": {
                "num": null,
                "type_str": "table",
                "text": "Work done at the University of Edinburgh.",
                "content": "<table><tr><td/><td/><td/><td>SBARQ</td><td/><td/><td/><td/></tr><tr><td colspan=\"9\">\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2534\u2500\u2500\u2500\u2500\u252c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510</td></tr><tr><td>\u2502</td><td/><td/><td/><td>SQ</td><td/><td/><td/><td>\u2502</td></tr><tr><td>\u2502</td><td/><td colspan=\"4\">\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2510</td><td/><td/><td>\u2502</td></tr><tr><td>\u2502</td><td/><td>\u2502</td><td/><td>\u2502</td><td>VP</td><td/><td/><td>\u2502</td></tr><tr><td>\u2502</td><td/><td>\u2502</td><td colspan=\"4\">\u250c\u2500\u2500\u2500 \u2502 \u2500\u2500\u2500\u2534\u2500\u2500\u2500\u2500\u2500\u2510</td><td/><td>\u2502</td></tr><tr><td>\u2502</td><td/><td>\u2502</td><td>VP</td><td>\u2502</td><td/><td>\u2502</td><td/><td>\u2502</td></tr><tr><td>\u2502</td><td colspan=\"7\">\u250c\u2500\u2500\u2500 \u2502 \u2500\u2500\u2500\u2534\u2500\u2500\u2500 \u2502 \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500 \u2502 \u2500\u2500\u2510</td><td>\u2502</td></tr><tr><td colspan=\"3\">\u2502 WHNP \u2502</td><td/><td>NP</td><td/><td>\u2502</td><td>\u2502</td><td>\u2502</td></tr><tr><td>\u2502</td><td>\u2502</td><td>\u2502</td><td colspan=\"3\">\u250c\u2500\u2500\u2500\u2500\u2534\u2500\u2500\u2500\u2500\u2510</td><td>\u2502</td><td>\u2502</td><td>\u2502</td></tr><tr><td colspan=\"3\">RB WP VBZ</td><td>DT</td><td/><td>NN</td><td colspan=\"3\">TO VB .</td></tr><tr><td>\u2502</td><td>\u2502</td><td>\u2502</td><td>\u2502</td><td/><td>\u2502</td><td>\u2502</td><td>\u2502</td><td>\u2502</td></tr><tr><td colspan=\"3\">So what 's</td><td>a</td><td/><td colspan=\"4\">parent to do ?</td></tr></table>",
                "html": null
            },
            "TABREF2": {
                "num": null,
                "type_str": "table",
                "text": "Full derivation for the sentence in",
                "content": "<table/>",
                "html": null
            },
            "TABREF3": {
                "num": null,
                "type_str": "table",
                "text": "Disc. F1 POS F1 Disc. F1 POS F1 Disc. F1 POS",
                "content": "<table><tr><td>static</td><td>91.1</td><td>68.2</td><td>97.2 87.4</td><td>61.7</td><td>98.3 83.6</td><td>51.3</td><td>97.9</td></tr><tr><td colspan=\"2\">dynamic 91.4</td><td>70.9</td><td>97.2 87.6</td><td>62.5</td><td>98.4 84.0</td><td>54.0</td><td>98.0</td></tr></table>",
                "html": null
            },
            "TABREF4": {
                "num": null,
                "type_str": "table",
                "text": "Results on development corpora. F1 is the Fscore on all constituents, Disc. F1 is an Fscore computed only on discontinuous constituents, POS is the accuracy on part-of-speech tags. Detailed results (including precision and recall) are given inTable 7of Appendix C.",
                "content": "<table><tr><td>English (DPTB) German (Tiger) German (Negra)</td></tr></table>",
                "html": null
            },
            "TABREF5": {
                "num": null,
                "type_str": "table",
                "text": "Discontinuous parsing results on the test sets.",
                "content": "<table/>",
                "html": null
            }
        }
    }
}