Abstract

Aspect-level sentiment analysis is a crucial problem in fine-grained sentiment analysis, which aims to automatically predict the sentiment polarity of the specific aspect in its context. In this talk, Dr. Yang will introduce several aspect-level sentiment analysis works: (i) feature enhanced attention network, which fully learns the connections among content words, aspect words and emotion words in sentences from different sub-spaces through multi-view mutual attention mechanism; (ii) hierarchical human-like strategy, which mimics the process of how humans identify the sentiment of a piece of text given a specific aspect; (iii) a large-scale multi-aspect multi-sentiment dataset, which would push forward the research of aspect-level sentiment analysis.