Research map / Vision-language and multimodal / Cross-modal learning
API: Adaptive Prototype Imputation for Incomplete Multimodal Sentiment Analysis
Xiaotao Wang, Yiyang Fang, Wenke Huang, Bin Yang, Guancheng Wan, Mang Ye
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 SCOPE class prototypes (static anchors) instead of zero-padding (ablation) | API · CMU-MOSI | F1 (%): 72.2 -> 76.9 | Improved |
| add DIAM instance-adaptive affine modulation on top of SCOPE (ablation) | API · CMU-MOSI | F1 (%): 76.9 -> 79.6 | Improved |
| add SCOPE prototypes (ablation) | API · CMU-MOSEI | F1 (%): 72.4 -> 74.0 | Improved |
| add DIAM on top of SCOPE (ablation) | API · CMU-MOSEI | F1 (%): 74.0 -> 77.8 | Improved |
| add SCOPE prototypes (ablation) | API · CMU-MOSEI | ACC2 (%): 73.4 -> 75.4 | Improved |
| add DIAM on top of SCOPE (ablation) | API · CMU-MOSEI | ACC2 (%): 75.4 -> 78.4 | Improved |
| API vs HME at average over missing rates | HME · CMU-MOSI | F1 (%): 78.7 -> 79.6 | Improved |
| API vs GCNet at extreme missingness MR=0.7 | GCNet · CMU-MOSI MR=0.7 | F1 (%): 65.4 -> 74.1 | Improved |
| API vs HME on MOSEI average | HME · CMU-MOSEI | F1 (%): 76.9 -> 77.8 | Improved |
| lambda_proto and lambda_rec set to 0.1 versus too small or too large (ablation) | API · CMU-MOSI 30% missing | ACC2 and F1 | Mixed |
| use API instead of IMDer for latency | IMDer · CMU-MOSI 1-Miss | latency (ms): 2241 -> 46 | Improved |
| use API instead of LNLN for missing-modality params | LNLN · CMU-MOSI | params (M): 2.99 -> 0.14 | Improved |
Closest papers by meaning
- A Unified Self-Distillation Framework for Multimodal Sentiment Analysis with Uncertain Missing Modalities (AAAI 2024 · similarity 0.89)
- SemMSA: Latent Semantic-Aided Robust Multimodal Sentiment Analysis with Incomplete Data (NeurIPS 2026 · similarity 0.88)
- Toward Robust Incomplete Multimodal Sentiment Analysis via Hierarchical Representation Learning (NeurIPS 2024 · similarity 0.88)
- Correlation-Decoupled Knowledge Distillation for Multimodal Sentiment Analysis with Incomplete Modalities (CVPR 2024 · similarity 0.87)
- A Multi-Focus-Driven Multi-Branch Network for Robust Multimodal Sentiment Analysis (AAAI 2025 · similarity 0.87)
- DLF: Disentangled-Language-Focused Multimodal Sentiment Analysis (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.
