Research map / Vision-language and multimodal / Cross-modal learning
On-the-Fly Modulation for Balanced Multimodal Learning
Yake Wei, Di Hu, Henghui Du, Ji-Rong Wen
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 OPM (adaptive modality dropout) | Concatenation fusion · CREMA-D | Accuracy: 66.9 -> 75.1 | Improved |
| add OGM (adaptive gradient modulation with noise) | Concatenation fusion · CREMA-D | Accuracy: 66.9 -> 74.6 | Improved |
| add OPM | Concatenation fusion · Kinetics-Sounds | Accuracy: 63.0 -> 67.0 | Improved |
| add OPM | FiLM · CREMA-D | Accuracy: 61.8 -> 62.4 | Improved |
| add OPM | VATT · Kinetics-Sounds | Accuracy: 61.8 -> 69.2 | Improved |
| add OPM | TSN · Kinetics-Sounds | Accuracy: 65.0 -> 70.0 | Improved |
| add OPM to multi-layer classifier model | Multi-layer classifier · CREMA-D | Accuracy: 67.7 -> 71.9 | Improved |
| add OPM to three-modality model | Concatenation · CMU-MOSI (audio+visual+text) | Accuracy: 75.9 -> 77.6 | Improved |
| replace Concatenation with OPM versus ModDrop fixed-probability modality dropout (ablation) | ModDrop · CREMA-D | Accuracy: 69.9 -> 75.1 | Improved |
| use standard neuron dropout instead of OPM (ablation) | Concatenation · CREMA-D | Accuracy: 66.9 -> 66.5 | No clear effect |
| combine OPM with OGM (ablation) | OPM · CREMA-D | Accuracy: 75.1 -> 76.7 | Improved |
| combine OPM with OGM (ablation) | OPM · VGGSound | Accuracy: 54.1 -> 53.3 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- FiLM — FILM: Following Instructions in Language with Modular Methods (ICLR 2022) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Classifier-guided Gradient Modulation for Enhanced Multimodal Learning (NeurIPS 2024 · similarity 0.90)
- Asymmetric Reinforcing Against Multi-Modal Representation Bias (AAAI 2025 · similarity 0.90)
- Modality-Aware SAM: Sharpness-Aware-Minimization Driven Gradient Modulation for Harmonized Multimodal Learning (NeurIPS 2025 · similarity 0.89)
- Improving Multimodal Learning Balance and Sufficiency through Data Remixing (ICML 2025 · similarity 0.89)
- Balancing Multimodal Training Through Game-Theoretic Regularization (NeurIPS 2025 · similarity 0.88)
- Multimodal Representation Learning by Alternating Unimodal Adaptation (CVPR 2024 · similarity 0.88)
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.
