The Ubuntu Dialogue Corpus: A Large Dataset for Research in Unstructured Multi-Turn Dialogue Systems
Ryan Lowe, Nissan Pow, Iulian Serban, Joelle Pineau
Code Available — Be the first to reproduce this paper.
ReproduceCode
- github.com/npow/ubottuOfficialIn papernone★ 0
- github.com/facebookresearch/ParlAIpytorch★ 10,627
- github.com/yf0726/SemeterProjectSpring19tf★ 0
- github.com/MichielDeRuiter/ConversationalClassificationtf★ 0
- github.com/david-yoon/QA_HRDE_LTCtf★ 0
- github.com/npow/ubuntu-corpusnone★ 0
- github.com/yangliuy/NeuralResponseRankingtf★ 0
- github.com/wayalhruhi/juliansernone★ 0
- github.com/dennybritz/chatbot-retrievaltf★ 0
- github.com/yangliuy/HybridNCMpytorch★ 0
Abstract
This paper introduces the Ubuntu Dialogue Corpus, a dataset containing almost 1 million multi-turn dialogues, with a total of over 7 million utterances and 100 million words. This provides a unique resource for research into building dialogue managers based on neural language models that can make use of large amounts of unlabeled data. The dataset has both the multi-turn property of conversations in the Dialog State Tracking Challenge datasets, and the unstructured nature of interactions from microblog services such as Twitter. We also describe two neural learning architectures suitable for analyzing this dataset, and provide benchmark performance on the task of selecting the best next response.
Tasks
Benchmark Results
| Dataset | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| Ubuntu Dialogue (v1, Ranking) | Dual-LSTM | R10@1 | 0.6 | — | Unverified |