Hybrid Uncertainty Calibration for Multimodal Sentiment Analysis
In open environments, multimodal sentiment analysis (MSA) often suffers from low-quality data and can be disrupted by noise, inherent defects, and outliers. In some cases, unreasonable multimodal fusion methods can perform worse than unimodal methods. Another challenge of MSA is effectively enabling...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-02-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/13/3/662 |