Topic-Conversation Relevance (TCR) Dataset and Benchmarks

Yaran Fan, Jamie Pool, Senja Filipi, Ross Cutler
2/3/2026

No official code found. Best implementation: microsoft/topic_conversation (10 stars, MIT, last commit 2026-02-09).

Abstract

Workplace meetings are vital to organizational collaboration, yet a large percentage of meetings are rated as ineffective. To help improve meeting effectiveness by understanding if the conversation is on topic, we create a comprehensive Topic-Conversation Relevance (TCR) dataset that covers a variety of domains and meeting styles. The TCR dataset includes 1,500 unique meetings, 22 million words in transcripts, and over 15,000 meeting topics, sourced from both newly collected Speech Interruption Meeting (SIM) data and existing public datasets. Along with the text data, we also open source scripts to generate synthetic meetings or create augmented meetings from the TCR dataset to enhance data diversity. For each data source, benchmarks are created using GPT-4 to evaluate the model accuracy in understanding transcription-topic relevance.

View on arXivView PDF

Code Implementations(1)

microsoft/topic_conversationCommunity100%
MIT

Topic-Conversation Relevance (TCR) Dataset and Benchmarks

102Python, Jupyter NotebookMay 22, 20248 months agoMIT

Similar Papers

A Survey on Recent Advances in Conversational Data Generation

77% match2024

You need to MIMIC to get FAME: Solving Meeting Transcript Scarcity with a Multi-Agent Conversations

77% match1 repo12025

Finding Diamonds in Conversation Haystacks: A Benchmark for Conversational Data Retrieval

77% match1 repo32025

DASH: Dialogue-Aware Similarity and Handshake Recognition for Topic Segmentation in Public-Channel Conversations

76% match1 repo42025

MEETING DELEGATE: Benchmarking LLMs on Attending Meetings on Our Behalf

76% match2025

Human Capital Visualization using Speech Amount during Meetings

76% match2025

Cite this paper

@article{fan2026topicconversation,
  title  = {Topic-Conversation Relevance (TCR) Dataset and Benchmarks},
  author = {Yaran Fan and Jamie Pool and Senja Filipi and Ross Cutler},
  year   = {2026},
  eprint = {2411.00038},
  archivePrefix = {arXiv},
  doi    = {10.48550/arXiv.2411.00038},
  url    = {https://arxiv.org/abs/2411.00038},
  journal = {NEURIPS 2024 2024}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2411.00038.svg)](https://codeofpaper.com/paper/2411.00038)

Discussion