Efficient Data Generation for Source-grounded Information-seeking Dialogs: A Use Case for Meeting Transcripts
This work addresses the challenge of automating data creation for information-seeking dialogs in specific domains like meeting analysis, offering an incremental improvement over manual methods.
The authors tackled the problem of generating data for source-grounded information-seeking dialogs over long, noisy meeting transcripts by proposing a semi-automatic LLM-based approach with human verification, resulting in the MISeD dataset that improves model performance, achieving comparable response generation to manual data while enhancing attribution and reducing effort.
Automating data generation with Large Language Models (LLMs) has become increasingly popular. In this work, we investigate the feasibility and effectiveness of LLM-based data generation in the challenging setting of source-grounded information-seeking dialogs, with response attribution, over long documents. Our source texts consist of long and noisy meeting transcripts, adding to the task complexity. Since automating attribution remains difficult, we propose a semi-automatic approach: dialog queries and responses are generated with LLMs, followed by human verification and identification of attribution spans. Using this approach, we created MISeD -- Meeting Information Seeking Dialogs dataset -- a dataset of information-seeking dialogs focused on meeting transcripts. Models finetuned with MISeD demonstrate superior performance compared to off-the-shelf models, even those of larger size. Finetuning on MISeD gives comparable response generation quality to finetuning on fully manual data, while improving attribution quality and reducing time and effort.