MediaSum: A Large-scale Media Interview Dataset for Dialogue Summarization
arXiv:2103.06410
Abstract
MediaSum, a large-scale media interview dataset consisting of 463.6K transcripts with abstractive summaries. To create this dataset, we collect interview transcripts from NPR and CNN and employ the overview and topic descriptions as summaries. Compared with existing public corpora for dialogue summarization, our dataset is an order of magnitude larger and contains complex multi-party conversations from multiple domains. We conduct statistical analysis to demonstrate the unique positional bias exhibited in the transcripts of televised and radioed interviews. We also show that MediaSum can be used in transfer learning to improve a model's performance on other dialogue summarization tasks.
arXiv commentDataset: https://github.com/zcgzcgzcg1/MediaSum/
References in corpus (7)
- Scikit-learn: Machine Learning in Python
- Unified Language Model Pre-training for Natural Language Understanding and Generation
- SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization
- Get To The Point: Summarization with Pointer-Generator Networks
- Abstractive Dialog Summarization with Semantic Scaffolds
- A Hierarchical Network for Abstractive Meeting Summarization with Cross-Domain Pretraining
- Interview: A Large-Scale Open-Source Corpus of Media Dialog