wise-east commited on
Commit
4755b63
1 Parent(s): c47b36b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +49 -15
README.md CHANGED
@@ -23,26 +23,41 @@ task_ids:
23
  - text-scoring
24
  - dialogue-modeling
25
  ---
26
- #### SPOLIN is provided in JSON format
 
27
  [![CC BY-NC 4.0][cc-by-nc-shield]][cc-by-nc]
28
 
29
 
30
- **First level keys:**
31
- * `yesands`
32
- * `non-yesands`
 
 
 
 
 
 
 
 
33
 
34
- **Second level keys:**
35
- * `spont`: shorthand for Spontaneanation
36
- * `cornell`: shorthand for Cornell Movie-Dialogs Corpus
37
- * `subtle`: (not present in `spolin-valid.json`) shorthand for SubTle Corpus
38
 
39
- Each second level item is a list of _yes-ands_ or non-_yes-ands_. Each item contains the following:
40
- * `id`: dataset ID
41
- * `prompt`: Utterance that starts the dialogue turn
42
- * `response`: Response to the prompt
43
 
 
 
 
 
 
 
 
 
 
44
 
45
- ##### `spolin-train.json`:
46
  || yesands| non-yesands|
47
  |--|---:|---:|
48
  |Spontaneanation|10,459|5,587*|
@@ -51,7 +66,7 @@ Each second level item is a list of _yes-ands_ or non-_yes-ands_. Each item cont
51
  |Total|67,188|43,409|
52
 
53
 
54
- ##### `spolin-train-acl.json`:
55
 
56
  || yesands| non-yesands|
57
  |--|---:|---:|
@@ -59,7 +74,7 @@ Each second level item is a list of _yes-ands_ or non-_yes-ands_. Each item cont
59
  |Cornell|14,976|17,851|
60
  |Total|25,435|23,438|
61
 
62
- ##### `spolin-valid.json`:
63
 
64
  || yesands| non-yesands|
65
  |--|---:|---:|
@@ -69,6 +84,25 @@ Each second level item is a list of _yes-ands_ or non-_yes-ands_. Each item cont
69
 
70
  \*Artificially collected by mix & matching positive Spontaneanation samples to balance dataset for training classifier
71
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
72
 
73
  ### License
74
 
 
23
  - text-scoring
24
  - dialogue-modeling
25
  ---
26
+
27
+ # SPOLIN
28
  [![CC BY-NC 4.0][cc-by-nc-shield]][cc-by-nc]
29
 
30
 
31
+ This is the repo for the paper ["Grounding Conversations with Improvised Dialogues"](https://aclanthology.org/2020.acl-main.218/) (ACL2020).
32
+ The _Selected Pairs of Learnable ImprovisatioN_ (SPOLIN) corpus is a collection of more than 68,000 "Yes, and" type dialogue pairs extracted from the Spontaneanation podcast by Paul F. Tompkins, the Cornell Movie-Dialogs Corpus, and the SubTle corpus. For more information, refer to our [paper](https://arxiv.org/abs/2004.09544) or our [project page](https://justin-cho.com/spolin).
33
+
34
+
35
+ ### Available SPOLIN versions:
36
+ The core dataset that was used for the experiments in the paper only includes _yes-ands_ and non-_yes-ands_ from Spontaneanation and most of what is provided in those extracted from the Cornell Movie-Dialogs Corpus. After the submitting the paper, we continued our iterative data augmentation process, repeating another iteration with the Cornell Movie-Dialogs Corpus and extracting from the SubTle corpus. This expanded version is also included in this repository [here](/data). This latest version of SPOLIN was used to train the model used in our [demo](https://spolin.isi.edu).
37
+
38
+
39
+ In the `data` folder, we provide two versions of the SPOLIN training set:
40
+ 1. Version used for experiments in the ACL paper: `data/spolin-train-acl.csv`
41
+ 2. Expanded version: `data/spolin-train.csv`
42
 
43
+ ### Relevant links:
 
 
 
44
 
45
+ * Project page: https://justin-cho.com/spolin
46
+ * Github repo: https://github.com/wise-east/spolin
47
+ * SpolinBot Demo: https://spolin.isi.edu
48
+ * ACL2020 Paper: https://aclanthology.org/2020.acl-main.218/
49
 
50
+ **Fiels**
51
+ * `id`: unique identifier
52
+ * `prompt`: first utterance in utterance pair
53
+ * `response`: second utterance in utterance pair
54
+ * `label`: yesand = 1, non-yesand = 0
55
+ * `source`: the source for the sample
56
+ * `split`: whether the sample belongs to the training set or the validation set
57
+
58
+ ### Dataset Statistics
59
 
60
+ ##### `spolin-train.csv`:
61
  || yesands| non-yesands|
62
  |--|---:|---:|
63
  |Spontaneanation|10,459|5,587*|
 
66
  |Total|67,188|43,409|
67
 
68
 
69
+ ##### `spolin-train-acl.csv`:
70
 
71
  || yesands| non-yesands|
72
  |--|---:|---:|
 
74
  |Cornell|14,976|17,851|
75
  |Total|25,435|23,438|
76
 
77
+ ##### `spolin-valid.csv`:
78
 
79
  || yesands| non-yesands|
80
  |--|---:|---:|
 
84
 
85
  \*Artificially collected by mix & matching positive Spontaneanation samples to balance dataset for training classifier
86
 
87
+ ### ACL Presentation
88
+
89
+ [Video recording](https://slideslive.com/38928948/grounding-conversations-with-improvised-dialogues)
90
+
91
+
92
+ ### Citation
93
+
94
+ If you use our data for your work, please cite our ACL2020 paper:
95
+ ```
96
+ @inproceedings{cho2020spolin,
97
+ title={Grounding Conversations with Improvised Dialogues},
98
+ author={Cho, Hyundong and May, Jonathan},
99
+ booktitle ={Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics},
100
+ publisher = {Association for Computational Linguistics},
101
+ location = {Seattle, Washington, USA},
102
+ year={2020}
103
+ }
104
+ ```
105
+
106
 
107
  ### License
108