Meta-based self-training and re-weighting for aspect-based sentiment analysis

Meta-based self-training and re-weighting for aspect-based sentiment analysis

Aspect-based sentiment analysis (ABSA) means to identify fine-grained aspects, opinions, and sentiment polarities. Recent ABSA research focuses on utilizing multi-task learning (MTL) to achieve less computational costs and better performance. However, there are certain limits in MTL-based ABSA. For...

Mô tả đầy đủ

Chi tiết về thư mục
Những tác giả chính: He, Kai, Mao, Rui, Gong, Tieliang, Li, Chen, Cambria, Erik
Tác giả khác: School of Computer Science and Engineering
Định dạng: Journal Article
Ngôn ngữ:English
Được phát hành: 2022
Những chủ đề:
Engineering::Computer science and engineering
Aspect-Based Sentiment Analysis
Meta Learning
Truy cập trực tuyến:https://hdl.handle.net/10356/163145

Internet

https://hdl.handle.net/10356/163145

Những quyển sách tương tự