File size: 11,076 Bytes
6fa4bc9 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 |
{
"paper_id": "1994",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T02:05:03.543320Z"
},
"title": "",
"authors": [
{
"first": "Jaime",
"middle": [],
"last": "Carbonell",
"suffix": "",
"affiliation": {},
"email": ""
},
{
"first": "David",
"middle": [],
"last": "Farwell",
"suffix": "",
"affiliation": {
"laboratory": "Computing Research Laboratory",
"institution": "Carnegie Mellon University New Mexico State University",
"location": {
"addrLine": "Schenley Park Box",
"postCode": "30001 / 3CRL, 15213-3890, 88003",
"settlement": "Pittsburgh, Las Cruces",
"region": "PA, NM"
}
},
"email": ""
},
{
"first": "Robert",
"middle": [],
"last": "Frederking",
"suffix": "",
"affiliation": {},
"email": ""
},
{
"first": "Steven",
"middle": [],
"last": "Helmreich",
"suffix": "",
"affiliation": {
"laboratory": "Computing Research Laboratory",
"institution": "Carnegie Mellon University New Mexico State University",
"location": {
"addrLine": "Schenley Park Box",
"postCode": "30001 / 3CRL, 15213-3890, 88003",
"settlement": "Pittsburgh, Las Cruces",
"region": "PA, NM"
}
},
"email": ""
},
{
"first": "Eduard",
"middle": [],
"last": "Hovy",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Southern California",
"location": {
"addrLine": "4676 Admiralty Way Marina del Rey",
"postCode": "90292-6695",
"region": "CA"
}
},
"email": ""
},
{
"first": "Kevin",
"middle": [],
"last": "Knight",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Southern California",
"location": {
"addrLine": "4676 Admiralty Way Marina del Rey",
"postCode": "90292-6695",
"region": "CA"
}
},
"email": ""
},
{
"first": "Lori",
"middle": [],
"last": "Levin",
"suffix": "",
"affiliation": {},
"email": ""
},
{
"first": "Sergei",
"middle": [],
"last": "Nirenburg",
"suffix": "",
"affiliation": {
"laboratory": "Computing Research Laboratory",
"institution": "Carnegie Mellon University New Mexico State University",
"location": {
"addrLine": "Schenley Park Box",
"postCode": "30001 / 3CRL, 15213-3890, 88003",
"settlement": "Pittsburgh, Las Cruces",
"region": "PA, NM"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "",
"pdf_parse": {
"paper_id": "1994",
"_pdf_hash": "",
"abstract": [],
"body_text": [
{
"text": "\u2022 The system contains three parallel translation engines for Spanish to English translation and two for Japanese to English translation. The three Spanish to English engines are:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "-lexical transfer (word-by-word and phrase-by-phrase substitution using a large bilingual glossary); -example-based MT (identification of phrases or even passages from the source text that appear in the large pre-constructed bilingual corpus of examples, and subsequent production of the translated phrases or passages as the target text); -knowledge-based MT (the more traditional route involving source text parsing, analysis, and generation, using an internal text representation that is gradually being upgraded to an Interlingua).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Each source text sentence is translated by all three engines, who produce one or more candidate translations plus rating for their goodness. A final translation fragment selection and text smoothing module integrates the best results from the various engines into a resulting target text.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "\u2022 Initially, the system was conceived as a human assistant, with gradual improvements and extensions until it would become the primary translator requiring only occasional human assistance and moderate postediting. A human interface called the Augmentor was built for use by translation assistants with a passing but not deep familiarity with the source languages. Due to contingencies of the ARPA evaluation paradigm, however, research effort has focused on the development of fully automated translation at significantly reduced quality.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "\u2022 The system involves a mixture of knowledge-based and statistical modules, as well as knowledge resources built up in various ways, by hand, from online dictionaries and knowledge bases, and through statistical extraction of knowledge from text.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "\u2022 The knowledge-based engine involves an Interlingua to represent the contents of the text being translated. Interlingua terms are defined in a taxonomy of approximately 80,000 concepts called the Ontology.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "\u2022 For parsing, CRL's parser PANGLYZER and Spanish grammar are used. The parser's output contains a mixture of syntactic and semantic information, following the theory of preference semantics. CRL is also responsible for the creation of the Spanish lexicon and the collection of other useful textual resources. The PANGLYZER is written in Quintus Prolog.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "\u2022 For the remaining semantic analysis and construction of the Interlingua statements corresponding to the input, software from ISI is used. The Ontology is being constructed at ISI with assistance from CRL and CMT.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "\u2022 For generation, ISI's PENMAN system is used. The software is written in Common Lisp.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "\u2022 CMT is responsible for the system architecture, the operator interface (including the Augmentor, the CMAT human-assistance editor, the inclusion of WordPerfect and Emacs text editing tools, etc.) and the successful integration of all the modules. CMT software is written in Common Lisp.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "\u2022 The Example-Based translation engine and resources are developed and built at CMT and CRL.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "\u2022 The lexical transfer engine and its bilingual glossary was developed at CMT.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Despite the differences in theoretical approach and implementation, our work on PANGLOSS makes clear that systems as different as the PANGLYZER, PENMAN, the various knowledge-based MT components at CMT can share knowledge resources in a meaningful way.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "This presentation will demonstrate the operation of the various engines of PANGLOSS on Spanish to English translation of newspaper texts. Since the complete system requires three Sun SPARC-Stations, only a scaled-down version of the system will be displayed. This version will however contain representative examples of the bilingual resources required by the EBMT and lexical transfer engines as well as all the intermediate steps produced in the chain of KBMT engine modules. The Ontology and various statistical routines will also be demonstrated. Finally, the operation of the CMAT editor and its function within the PANGLOSS Translator's Workstation will be shown.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "The Presentation",
"sec_num": null
}
],
"back_matter": [],
"bib_entries": {},
"ref_entries": {}
}
} |