Research map / Vision-language and multimodal / Cross-modal learning
Supporting Multimodal Intermediate Fusion with Informatic Constraint and Distribution Coherence
Yi Li, Fei Song, Changwen Zheng, Jiangmeng 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 |
|---|---|---|---|
| replace late fusion with intermediate fusion | QMF · MVSA-Single | accuracy (%): 78.07 -> 78.81 | Improved |
| replace late fusion with intermediate fusion | PDF · MVSA-Single | accuracy (%): 79.94 -> 80.36 | Improved |
| add distribution cohering module (ablation) | IID-P · MVSA-Single | accuracy (%): 80.79 -> 81.13 | Improved |
| add informatic constraint (ablation) | IID-P · MVSA-Single | accuracy (%): 80.46 -> 81.13 | Improved |
| add distribution cohering module (ablation) | IID-L · MVSA-Single | accuracy (%): 77.42 -> 77.78 | Improved |
| add informatic constraint (ablation) | IID-Q · MVSA-Single | accuracy (%): 78.93 -> 80.02 | Improved |
| add informatic constraint (ablation) | IID-P · Food101 | accuracy (%): 93.60 -> 93.73 | Improved |
| add distribution cohering module (ablation) | IID-P · SUN RGB-D | accuracy (%): 62.77 -> 62.99 | Improved |
| plug IID into OTKGE | OTKGE · FB-IMG | MRR: 0.843 -> 0.855 | Improved |
| plug IID into MMKRL | MMKRL · WN9-IMG | MRR: 0.913 -> 0.920 | Improved |
| add IID to PDF fusion weights | PDF · MVSA-Multiple | accuracy (%): 69.54 -> 71.23 | Improved |
Closest papers by meaning
- Unity by Diversity: Improved Representation Learning for Multimodal VAEs (NeurIPS 2024 · similarity 0.88)
- Predictive Dynamic Fusion (ICML 2024 · similarity 0.87)
- Towards the Causal Complete Cause of Multi-Modal Representation Learning (ICML 2025 · similarity 0.87)
- A Closer Look at Multimodal Representation Collapse (ICML 2025 · similarity 0.87)
- Amplifying Prominent Representations in Multimodal Learning via Variational Dirichlet Process (NeurIPS 2025 · similarity 0.87)
- Aligning Multimodal Representations through an Information Bottleneck (ICML 2025 · similarity 0.87)
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.
