Tri-MCA fusion: cross-modal attention and dynamic gating for multimodal sentiment analysis
Asmaa Alrayzah · 2026
Read Tri-MCA fusion: cross-modal attention and dynamic gating for multimodal sentiment analysis by Asmaa Alrayzah (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13470083.txt.
Description
Asmaa Alrayzah built a tri-modal sentiment model in 2026 that fuses text, voice, and video with cross-attention and adaptive gating. Single-mode sentiment misses tone and expression, while static fusion mishandles modality imbalance and noise. The Tri-MCA design lets each modality attend to the others, with gates weighting their contributions per case.
On CMU-MOSI the model scored 85.6% binary accuracy and 85.2 F1, beating several rivals, with similar gains on CMU-MOSEI and SIMS. Ablation tests confirmed each component's contribution, especially the pairing of cross-attention with gating. The framework offers a robust route to emotion reading for assistants, media analysis, and human-computer interaction.
Catalog record
| CREATOR | Asmaa Alrayzah |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 70 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |