Multi-Head Self-Attention Transformation Networks for Aspect-Based Sentiment Analysis

Multi-Head Self-Attention Transformation Networks for Aspect-Based Sentiment Analysis

Aspect-based sentiment analysis (ABSA) aims to analyze the sentiment polarity of an input sentence in a certain aspect. Many existing methods of ABSA employ long short-term memory (LSTM) networks and attention mechanism. However, the attention mechanism only models the local certain dependencies of...

Full description

Bibliographic Details
Main Authors: Yuming Lin, Chaoqiang Wang, Hao Song, You Li
Format: Article
Language:English
Published: IEEE 2021-01-01
Series:IEEE Access
Subjects:
Aspect-based sentiment analysis
self-attention
transformation networks
target-specific transformation
Online Access:https://ieeexplore.ieee.org/document/9314135/

Internet

https://ieeexplore.ieee.org/document/9314135/

Similar Items