Skip to main navigation Skip to search Skip to main content

Understanding Attention for Vision-and-Language Tasks

Research output: Chapter in Book/Conference paper › Conference paper › peer-review

Abstract

Attention mechanism has been used as an important component across Vision-and-Language(VL) tasks in order to bridge the semantic gap between visual and textual features. While attention has been widely used in VL tasks, it has not been examined the capability of different attention alignment calculation in bridging the semantic gap between visual and textual clues. In this research, we conduct a comprehensive analysis on understanding the role of attention alignment by looking into the attention score calculation methods and check how it actually represents the visual region’s and textual token’s significance for the global assessment. We also analyse the conditions which attention score calculation mechanism would be more (or less) interpretable, and which may impact the model performance on three different VL tasks, including visual question answering, text-to-image generation, text-and-image matching (both sentence and image retrieval). Our analysis is the first of its kind and provides useful insights of the importance of each attention alignment score calculation when applied at the training phase of VL tasks, commonly ignored in attention-based cross modal models, and/or pretrained models. Our code is available at: https://github.com/adlnlp/Attention_VL
Original languageEnglish
Title of host publicationProceedings of the 29th International Conference on Computational Linguistics
Place of PublicationGyeongju, Republic of Korea
PublisherInternational Committee on Computational Linguistics
Pages3438-3453
Number of pages16
Volume29
Edition1
Publication statusPublished - 2022
EventThe 29th International Conference on Computational Linguistics - Gyeongju, Korea, Republic of
Duration: 12 Oct 2022 → 17 Oct 2022
Conference number: 29
https://coling2022.org/

Publication series

NameProceedings - International Conference on Computational Linguistics, COLING
ISSN (Print)2951-2093

Conference

ConferenceThe 29th International Conference on Computational Linguistics
Abbreviated titlecoling2022
Country/TerritoryKorea, Republic of
CityGyeongju
Period12/10/22 → 17/10/22
Internet address

Fingerprint

Dive into the research topics of 'Understanding Attention for Vision-and-Language Tasks'. Together they form a unique fingerprint.

Cite this