Research Seminar by Jianfei Yu | Multimodal Entity-Based Sentiment Analysis on Social Media

Multimodal Entity-Based Sentiment Analysis on Social Media

As interpersonal communications in multimedia have become increasingly popular, more and more users on social platforms such as Twitter and Facebook are turning to post multimedia content with paired text and image to share personal ideas and opinions. To identify the public opinions expressed in these massive multimodal posts, multimodal entity-based sentiment analysis has recently attracted considerable attention, and significant progress has been made owing to the fast development of pre-trained models and large multimodal models. In this talk, I will first introduce the background and subtasks of multimodal entity-based sentiment analysis, and then summarize our several recent ACL and MM papers on multimodal named entity recognition and entity-based sentiment analysis. Moreover, I will briefly talk about our recent works in other related directions, such as Multimodal Emotion Cause Analysis in Conversations.