arrow
返回

A Deep Multi-task Contextual Attention Framework for Multi-modal Affect Analysis

delete2020-05-13
delete29
PRE
AI
M
Md Shad Akhtar *
D
Dushyant Singh Chauhan
A
Asif Ekbal
DOI:10.1145/3380744delete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
Multi-modal affect analysis (e.g., sentiment and emotion analysis) is an interdisciplinary study and has been an emerging and prominent field in Natural Language Processing and Computer Vision. The effective fusion of multiple modalities (e.g., text, acoustic, or visual frames) is a non-trivial task, as these modalities, often, carry distinct and diverse information, and do not contribute equally. The issue further escalates when these data contain noise. In this article, we study the concept of multi-task learning for multi-modal affect analysis and explore a contextual inter-modal attention framework that aims to leverage the association among the neighboring utterances and their multi-modal information. In general, sentiments and emotions have inter-dependence on each other (e.g., anger. negative or happy. positive). In our current work, we exploit the relatedness among the participating tasks in the multi-task framework. We define three different multi-task setups, each having two tasks, i.e., sentiment & emotion classification, sentiment classification & sentiment intensity prediction, and emotion classification & emotion intensity prediction. Our evaluation of the proposed system on the CMU-Multi-modal Opinion Sentiment and Emotion Intensity benchmark dataset suggests that, in comparison with the single-task learning framework, our multi-task framework yields better performance for the inter-related participating tasks. Further, comparative studies show that our proposed approach attains state-of-the-art performance for most of the cases.
Keyword:
Multi-task learning
multi-modal analysis
sentiment analysis
sentiment intensity prediction
emotion analysis
emotion intensity prediction
inter-modal attention
AI总结

AI总结

对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。

期刊

ACM Transactions on Knowledge Discovery from Data 封面图
ACM Transactions on Knowledge Discovery from Data
IF:
4.8
论文数:
1.3K
被引数:
4.4K

机构

I
indian institute of technology system (iit system)
学者数:
9.5W
论文数: 9.9W
被引数: 93
I
Indraprastha Institute of Information Technology Delhi
学者数:
933
论文数: 689
被引数: 558
引用论文

引用论文

End-to-End Multimodal Emotion Recognition Using Deep Neural Networks基于深度神经网络的端到端多模态情感识别
err2017-12-01
err375
errOAAI
errTzirakis, Panagiotis; Trigeorgis, George; Nicolaou, Mihalis A.; Schuller, Bjorn W.; Zafeiriou, Stefanos
err分享
err收藏
Aspect extraction for opinion mining with a deep convolutional neural network
err2016-09-01
err600
PREAI
errPoria, Soujanya; Cambria, Erik; Gelbukh, Alexander
err分享
err收藏
Zolpidem is a potent anticonvulsant in adult and aged mice
err2010-01-01
err0
PREAI
errJosipa Vlainić; Danka Peričić
err分享
err收藏
5-farnesyloxycoumarin: a potent 15-LOX-1 inhibitor, prevents prostate cancer cell growth
err2016-11-15
err0
PREAI
errAla Orafaie; Hamid Sadeghian; Ahmad Reza Bahrami; Saffiyeh Saboormaleki; Maryam M. Matin
err分享
err收藏
The Physiology of Orthostatic Tremor
err1986-06-01
err0
PREAI
errP. D. Thompson; J. C. Rothwell; B. L. Day; A. Berardelli; J. P. R. Dick; T. Kachi; C. D. Marsden
err分享
err收藏
Androgen Action in Prostate Cancer
err2010-11-16
err0
errOAAI
errSujit Basu; Donald J. Tindall
err分享
err收藏
err分享
err收藏
学者 查看更多内容