Papers with multimodal sentiment analysis models
Sentiment Word Aware Multimodal Refinement for Multimodal Sentiment Analysis with ASR Errors (2022.findings-acl)
Copied to clipboard
| Challenge: | Existing models for multimodal sentiment analysis are limited in their capacity to be deployed in the real world. |
| Approach: | They propose a model that can dynamically refine erroneous sentiment words by leveraging multimodal sentiment clues. |
| Outcome: | The proposed model surpasses the state-of-the-art models on three datasets. |