Jotter: An Approach to Summarize the Formal Online Meeting

This paper examines various meeting summarising techniques. It also presents a hybridised technique for meeting summarization that combines abstractive and extractive techniques. Text extraction from meetings is accomplished in two ways: audio during the meeting and text recognition from the screen. These extracted texts are then compiled to create a meeting summary. In contrast to extractive text summarising, which concatenates essential sentences from paragraphs, abstractive text summarization focuses on providing a coherent summary of the given content. The maj ority of recent progress on abstractive summarization has been using RNN or Recurrent Neural Networks, although RNN-based algorithms fail when dealing with long sequences. This paper proposes hybridizing these two methods sequentially for meeting summarization to create a better overall summary of the meeting. This research endeavour seeks to convert conference video and audio into text documents and then run our hybridised text summarizer model on them. The result is a summarised text document that the various stakeholders, whether present at the conference or not, could utilise for rapid and concise reference. It will help the people who are not been able to attend the meeting can get the main points discussed in the meeting through summary.