Summary of Spontaneous Informal Speech Dataset For Punctuation Restoration, by Xing Yi Liu and Homayoon Beigi
Spontaneous Informal Speech Dataset for Punctuation Restoration
by Xing Yi Liu, Homayoon Beigi
First submitted to arxiv on: 17 Sep 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Human-Computer Interaction (cs.HC); Machine Learning (cs.LG); Sound (cs.SD); Audio and Speech Processing (eess.AS)
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 The paper introduces SponSpeech, a punctuation restoration dataset derived from informal speech sources, to address the discrepancy in evaluating punctuation models on scripted corpora versus real-world spontaneous speech. The dataset includes punctuation and casing information, as well as a filtering pipeline that examines audio and transcription quality. Additionally, the authors construct a challenging test set to evaluate models’ ability to leverage audio information to predict grammatically ambiguous punctuation. The paper contributes a publicly available dataset and code for building the dataset and running model experiments. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Punctuation restoration models are usually tested on well-structured speech, but real-world systems face messy, unpredictable speech with mistakes. To fix this, researchers created SponSpeech, a big collection of imperfect speech that includes the correct punctuation. They also built a tool to make more data and tested their models on tricky cases where audio clues help predict punctuation. |
