< Back
Posted: 20 Aug 2021 21:00

“Dialogue Summarization” August 2021 — summary from Astrophysics Data System

Brevi Assistant
Brevi Assistant

Business performance assistant

“Dialogue Summarization” August 2021 — summary from Astrophysics Data System main image

With the growth of dialogue systems and all-natural language generation methods, the resurgence of dialogue summarization has attracted significant study interests, which intends to condense the initial dialogue into a shorter version covering salient details. We wish that this first study of dialogue summarization can provide the community with a fast gain access to and a general photo to this job and inspire future researches. In this paper, we aim to enhance abstractive dialogue summarization top quality and, at the very same time, allow granularity control. A simple method to control the granularity of the final summary, in that our version can instantly figure out or regulate the number of generated recap sentences for a provided dialogue by anticipating and highlighting different text extends from the source text. Summarizing conversations via neural approaches has been acquiring research traction recently, yet it is still challenging to acquire practical options. As a result, in this work, we examine different techniques to explicitly include coreference information in neural abstractive dialogue summarization versions to tackle the abovementioned challenges. Dialogue summarization aims to generate a summary that suggests the bottom lines of a given dialogue. The fact regularization motivates the produced recap to be factually consistent with the ground-truth summary throughout version training, which helps to enhance the accurate accuracy of the created summary in inference time. Current dialogue summarization systems typically inscribe the message with a variety of basic semantic attributes e. g., topics and keyword phrases to gain a lot more powerful dialogue modeling capabilities. In this paper, we reveal exactly how DialoGPT, a pre-trained design for conversational response generation, can be developed as an unsupervised dialogue annotator, which benefits from dialogue background expertise inscribed in DialoGPT.


Source texts:


logo

The Brevi assistant is a novel way to summarize, assemble, and consolidate multiple text documents, reports, reviews, feedback, etc.

Partners:

© All rights reserved
2021 made by Brevi Technologies Inc.