Summary of Augsumm: Towards Generalizable Speech Summarization Using Synthetic Labels From Large Language Model, by Jee-weon Jung et al.
AugSumm: towards generalizable speech summarization using synthetic labels from large language model
by Jee-weon Jung, Roshan Sharma, William Chen, Bhiksha Raj, Shinji Watanabe
First submitted to arxiv on: 10 Jan 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a novel method for abstractive speech summarization (SSUM) that generates human-like summaries from speech recordings. The challenge in training SSUM models is that conventional approaches rely on a single ground-truth (GT) summary, which may not capture the variability in information and phrasing present in recordings. To address this, the authors introduce AugSumm, a method that leverages large language models (LLMs) to generate synthetic summaries, which can be used as proxies for human annotators. The paper explores different prompting strategies to generate these synthetic summaries using ChatGPT and validates their quality through human evaluation. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research aims to improve abstractive speech summarization by generating multiple possible summaries instead of just one. To do this, they use large language models like ChatGPT to create extra summaries that can be used to train and test the summarization models. The authors tested these new methods on a dataset called How2 and found that using the synthetic summaries helped improve the accuracy of the summarizations. |
Keywords
» Artificial intelligence » Prompting » Summarization
