A Sliding-Window Approach to Automatic Creation of Meeting Minutes

Abstract

Meeting minutes record any subject matters discussed, decisions reached and actions taken at meetings. The importance of minuting cannot be overemphasized in a time when a significant number of meetings take place in the virtual space. In this paper, we present a sliding window approach to automatic generation of meeting minutes. It aims to tackle issues associated with the nature of spoken text, including lengthy transcripts and lack of document structure, which make it difficult to identify salient content to be included in the meeting minutes. Our approach combines a sliding window and a neural abstractive summarizer to navigate through the transcripts to find salient content. The approach is evaluated on transcripts of natural meeting conversations, where we compare results obtained for human transcripts and two versions of automatic transcripts and discuss how and to what extent the summarizer succeeds at capturing salient content.

Semi-automatic Summarization of Spoken Discourse for Recording Ideas using GPT-3

“…Although there exist several studies about summarization of spoken meeting records using automatic speech recognition, human-machine collaboration we are focusing on has not been considered. For example, Song et al proposed a CNN-based system for automatic meeting transcription, summarization called SmartMeeting [4], Koay et al proposed a BART-based summarization system with a sliding-window approach [5] and Alexandra et al proposed a phrasal query-based summarization system with BART [6]. Although these studies dealt with similar target to our one, they did not dealt with humanmachine collaborative summarization of spoken discourse.…”