mathiascreutz commited on
Commit
ebea69a
1 Parent(s): 14e8e04

Minor modifications

Browse files
Files changed (1) hide show
  1. README.md +19 -3
README.md CHANGED
@@ -81,17 +81,33 @@ task_ids:
81
 
82
  - **Paper:** [Mathias Creutz (2018): Open Subtitles Paraphrases Corpus For Six Languages](http://www.lrec-conf.org/proceedings/lrec2018/summaries/131.html)
83
 
84
- - **Point of Contact:** [Mathias Creutz](firstname dot lastname at helsinki dot fi)
85
 
86
  ### Dataset Summary
87
 
88
- TBA
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
89
 
90
  ### Supported Tasks and Leaderboards
91
 
92
  **Tasks:** Paraphrase detection and generation
93
 
94
- Currently there is no Leaderboard for this dataset.
95
 
96
  ### Languages
97
 
 
81
 
82
  - **Paper:** [Mathias Creutz (2018): Open Subtitles Paraphrases Corpus For Six Languages](http://www.lrec-conf.org/proceedings/lrec2018/summaries/131.html)
83
 
84
+ - **Point of Contact:** Mathias Creutz (firstname dot lastname at helsinki dot fi)
85
 
86
  ### Dataset Summary
87
 
88
+ Opusparcus is a paraphrase corpus for six European languages: German,
89
+ English, Finnish, French, Russian, and Swedish. The paraphrases are
90
+ extracted from the OpenSubtitles2016 corpus, which contains subtitles
91
+ from movies and TV shows.
92
+
93
+ The data in Opusparcus has been extracted from OpenSubtitles2016
94
+ (http://opus.nlpl.eu/OpenSubtitles2016.php), which is in turn based on
95
+ data from http://www.opensubtitles.org/.
96
+
97
+ For each target language, the Opusparcus data have been partitioned
98
+ into three types of data sets: training, validation and test sets. The
99
+ training sets are large, consisting of millions of sentence pairs, and
100
+ have been compiled automatically, with the help of probabilistic
101
+ ranking functions. The development and test sets consist of sentence
102
+ pairs that have been annotated manually; each set contains
103
+ approximately 1000 sentence pairs that have been verified to be
104
+ acceptable paraphrases by two indepedent annotators.
105
 
106
  ### Supported Tasks and Leaderboards
107
 
108
  **Tasks:** Paraphrase detection and generation
109
 
110
+ **Leaderboards:** Currently there is no Leaderboard for this dataset.
111
 
112
  ### Languages
113