SANVis: Visual Analytics for Understanding Self-Attention Networks

Cheonbok Park, Jaegul Choo, Inyoup Na, Yongjang Jo, Sungbok Shin, Jaehyo Yoo, Bum Chul Kwon, Jian Zhao, Hyungjong Noh, Yeonsoo Lee

Research output: Chapter in Book/Report/Conference proceedingConference contribution

11 Citations (Scopus)

Abstract

Attention networks, a deep neural network architecture inspired by humans' attention mechanism, have seen significant success in image captioning, machine translation, and many other applications. Recently, they have been further evolved into an advanced approach called multi-head self-attention networks, which can encode a set of input vectors, e.g., word vectors in a sentence, into another set of vectors. Such encoding aims at simultaneously capturing diverse syntactic and semantic features within a set, each of which corresponds to a particular attention head, forming altogether multi-head attention. Meanwhile, the increased model complexity prevents users from easily understanding and manipulating the inner workings of models. To tackle the challenges, we present a visual analytics system called SANVis, which helps users understand the behaviors and the characteristics of multi-head self-attention networks. Using a state-of-the-art self-attention model called Transformer, we demonstrate usage scenarios of SANVis in machine translation tasks. Our system is available at http://short.sanvis.org.

Original languageEnglish
Title of host publication2019 IEEE Visualization Conference, VIS 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages146-150
Number of pages5
ISBN (Electronic)9781728149417
DOIs
Publication statusPublished - 2019 Oct
Event2019 IEEE Visualization Conference, VIS 2019 - Vancouver, Canada
Duration: 2019 Oct 202019 Oct 25

Publication series

Name2019 IEEE Visualization Conference, VIS 2019

Conference

Conference2019 IEEE Visualization Conference, VIS 2019
Country/TerritoryCanada
CityVancouver
Period19/10/2019/10/25

Keywords

  • Deep neural networks
  • interpretability
  • natural language processing
  • self-attention networks
  • visual analytics

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Media Technology
  • Modelling and Simulation

Fingerprint

Dive into the research topics of 'SANVis: Visual Analytics for Understanding Self-Attention Networks'. Together they form a unique fingerprint.

Cite this