Automatic Minuting: A Pipeline Method for Generating Minutes from Multi-Party Meeting Transcripts

Automatically generating meeting minutes is a challenging yet time-relevant problem in speech and natural language processing. Nowadays, meeting minutes seem more crucial than ever due to the manifold rise of online meetings. However, automatic minuting is not straightforward for various reasons: obtaining transcriptions of sufficient quality, summarizing long dialogue discourse, retaining topical relevance and coverage, handling redundancies and small talk, etc. This paper presents our investigations on a pipelined approach to automatically generate meeting minutes using a BART model (Bidirectional and Auto-Regressive Transformers) trained on multi-party dialogue summarization datasets. We achieve comparable results with our simple yet intuitive method with respect to previous large and computationally heavy state-of-the-art models. We make our code available at https://github.com/ELITR/ minuting-pipeline.