A Survey of Cross-Lingual Sentiment Analysis Based on Pre-Trained Models

Authors

Xudong Luo, Guangxi Key Lab of Multi-Source Information Mining & Security, Guangxi Normal University, Guilin, ChinaFollow
Sihan Yin, Guangxi Key Lab of Multi-Source Information Mining & Security, Guangxi Normal University, Guilin, ChinaFollow
Pingping Lin, Guangxi Key Lab of Multi-Source Information Mining & Security, Guangxi Normal University, Guilin, ChinaFollow

Abstract

With the technology development of natural language processing, many researchers have studied Machine Learning (ML), Deep Learning (DL), monolingual Sentiment Analysis (SA) widely. However, there is not much work on Cross-Lingual SA (CLSA), although it is beneficial when dealing with low resource languages (e.g., Tamil, Malayalam, Hindi, and Arabic). This paper surveys the main challenges and issues of CLSA based on some pre-trained language models and mentions the leading methods to cope with CLSA. In particular, we compare and analyze their pros and cons. Moreover, we summarize the valuable cross-lingual resources and point out the main problems researchers need to solve in the future.

Recommended Citation

Luo, Xudong; Yin, Sihan; and Lin, Pingping, "A Survey of Cross-Lingual Sentiment Analysis Based on Pre-Trained Models" (2021). ICEB 2021 Proceedings (Nanjing, China). 18.
https://aisel.aisnet.org/iceb2021/18