{"title":"Team The Turing TESTament @ AutoMin 2021: A Pipeline based Approach to Generate Meeting Minutes Using TOPSIS","authors":"Umang Sharma, Muskaan Singh, Harpreet Singh","doi":"10.21437/automin.2021-9","DOIUrl":null,"url":null,"abstract":"In this paper, we present our submission for AutoMin Shared Task@INTERSPEECH 2021. The objectives in this task were divided into three tasks, with the main task to create a summary based on a transcript from a meeting. The other two tasks were to compare minutes and transcripts to find out if they were from the same meeting or not. We propose a pipeline-based system that extracts the important sentences from the transcript using features and then a topsis algorithm to summarize. It cre-ates a flexible system that can provide a set of sentences from any given transcript that can best describe it based on selected features and heuristic evaluation metrics. The proposed system presents readable, grammatically correct, and fluent minutes for given meeting transcripts. We make our codebase ac-cessible here https://github.com/umangSharmacs/ theTuringTestament .","PeriodicalId":186820,"journal":{"name":"First Shared Task on Automatic Minuting at Interspeech 2021","volume":"17 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"First Shared Task on Automatic Minuting at Interspeech 2021","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.21437/automin.2021-9","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
Abstract
In this paper, we present our submission for AutoMin Shared Task@INTERSPEECH 2021. The objectives in this task were divided into three tasks, with the main task to create a summary based on a transcript from a meeting. The other two tasks were to compare minutes and transcripts to find out if they were from the same meeting or not. We propose a pipeline-based system that extracts the important sentences from the transcript using features and then a topsis algorithm to summarize. It cre-ates a flexible system that can provide a set of sentences from any given transcript that can best describe it based on selected features and heuristic evaluation metrics. The proposed system presents readable, grammatically correct, and fluent minutes for given meeting transcripts. We make our codebase ac-cessible here https://github.com/umangSharmacs/ theTuringTestament .