Research map / Vision-language and multimodal / Cross-modal learning
Beyond Missing Modalities: Hypergraph Conditioned Diffusion for Uncertainty-Aware Multimodal Emotion Recognition
Xihang Qiu, Yuhao Fang, Qing Zhou, Bin Zhai, Jialong Hong, Wanpeng Zhang, Yao Lu, Ye Zhang, Chun Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add KL regularizer on top of vacuity loss (ablation) | HyperEF · IEMOCAP4 missing rate 0.3 | accuracy (%): 48.75 -> 80.31 | Improved |
| add vacuity loss on top of KL regularizer (ablation) | HyperEF · MELD missing rate 0.3 | accuracy (%): 50.40 -> 63.78 | Improved |
| add vacuity loss alone to cross-entropy (ablation) | HyperEF · IEMOCAP4 missing rate 0.3 | accuracy (%): 76.85 -> 48.75 | Hurt |
| add KL regularizer alone to cross-entropy (ablation) | HyperEF · MELD missing rate 0.3 | accuracy (%): 55.06 -> 50.40 | Hurt |
| pair source-level uncertainty with entropy (discriminative level) rather than vacuity, dissonance or consonance (ablation) | DCEF · MELD missing rate 0.3 | accuracy (%): 60.28 -> 63.78 | Improved |
| pair source-level uncertainty with vacuity as second axis (ablation) | DCEF · MELD missing rate 0.3 | accuracy (%): 60.28 -> 59.87 | Hurt |
| use MHGAT instead of 2-layer Transformer encoder (ablation) | HyperEF · IEMOCAP4 missing rate 0.5 | accuracy (%): 76.2 -> 77.1 | Improved |
| use MHGAT instead of Transformer encoder (ablation) | HyperEF · MDLE missing rate 0.3 | accuracy (%): 62.5 -> 62.4 | No clear effect |
| MHGAT-conditioned diffusion recovery vs score-based IMDer | HyperEF · MELD feature recovery (text) | MSE: 0.701 -> 0.028 | Improved |
| HyperEF vs SDR-GNN across missing rates | HyperEF · IEMOCAP6 missing rate 0.6 | accuracy (%): 55.5 -> 59.5 | Improved |
Closest papers by meaning
- Federated Dialogue-Semantic Diffusion for Emotion Recognition under Incomplete Modalities (NeurIPS 2025 · similarity 0.90)
- Dynamic Interactive Bimodal Hypergraph Networks for Emotion Recognition in Conversations (AAAI 2025 · similarity 0.88)
- Adaptive Graph Learning for Multimodal Conversational Emotion Detection (AAAI 2024 · similarity 0.88)
- Multi-to-Single: Reducing Multimodal Dependency in Emotion Recognition Through Contrastive Learning (AAAI 2025 · similarity 0.87)
- BIG-FUSION: Brain-Inspired Global-Local Context Fusion Framework for Multimodal Emotion Recognition in Conversations (AAAI 2025 · similarity 0.86)
- Revisiting Multimodal Emotion Recognition in Conversation from the Perspective of Graph Spectrum (AAAI 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
