Summary of Topic-conversation Relevance (tcr) Dataset and Benchmarks, by Yaran Fan et al.
Topic-Conversation Relevance (TCR) Dataset and Benchmarks
by Yaran Fan, Jamie Pool, Senja Filipi, Ross Cutler
First submitted to arxiv on: 29 Oct 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 Medium Difficulty summary: This paper tackles the issue of ineffective workplace meetings by developing a comprehensive Topic-Conversation Relevance (TCR) dataset. The TCR dataset comprises 1,500 unique meetings, spanning various domains and meeting styles, with over 22 million words in transcripts and more than 15,000 meeting topics. The dataset combines newly collected SIM data with existing public datasets. To enhance diversity, the paper also provides open-source scripts for generating synthetic or augmented meetings. For each data source, benchmarks are created using GPT-4 to evaluate the model’s accuracy in understanding transcription-topic relevance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Low Difficulty summary: Meetings at work are important, but many aren’t very effective. This research aims to make meetings better by creating a big dataset of conversations and topics. The dataset has 1,500 different meetings, with over 22 million words spoken and more than 15,000 topics discussed. The data comes from new recordings and existing datasets. To make the data even more useful, the researchers also provide tools to create fake or modified meeting conversations. They used a special AI model called GPT-4 to test how well it can understand what’s being talked about during meetings. |
Keywords
» Artificial intelligence » Gpt
