Bisi commited on
Commit
99e00aa
1 Parent(s): ef3e47b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +35 -1
README.md CHANGED
@@ -3,4 +3,38 @@ task_categories:
3
  - summarization
4
  - text-generation
5
  - text2text-generation
6
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  - summarization
4
  - text-generation
5
  - text2text-generation
6
+ ---
7
+
8
+ # DivSumm summarization dataset
9
+
10
+ Dataset introduced in the paper: Analyzing the Dialect Diversity in Multi-document Summaries (COLING 2022)
11
+ _Olubusayo Olabisi, Aaron Hudson, Antonie Jetter, Ameeta Agrawal_
12
+
13
+
14
+ DivSumm is a novel dataset consisting of dialect-diverse tweets and human-written extractive and abstractive summaries. It consists of 90 tweets each on 25 topics in multiple English dialects (African-American, Hispanic and White), and two reference summaries per input.
15
+
16
+
17
+
18
+ ## Directories
19
+
20
+ input_docs - 90 tweets per topic evenly distributed among 3 dialects; total 25 topics
21
+
22
+ abstractive - Two annotators were asked to summarize each topic in 5 sentences using their own words.
23
+
24
+ extractive - Two annotators were asked to select 5 tweets from each topic that summarized the input tweets.
25
+
26
+
27
+ ## Paper
28
+
29
+
30
+ You can find our paper [here](https://aclanthology.org/2022.coling-1.542/). If you use this dataset in your work, please cite our paper:
31
+
32
+
33
+ @inproceedings{olabisi-etal-2022-analyzing,
34
+ title = "Analyzing the Dialect Diversity in Multi-document Summaries",
35
+ author = "Olabisi, Olubusayo and Hudson, Aaron and Jetter, Antonie and Agrawal, Ameeta",
36
+ booktitle = "Proceedings of the 29th International Conference on Computational Linguistics",
37
+ month = oct,
38
+ year = "2022",
39
+ }
40
+