Research map / Vision-language and multimodal / Cross-modal learning
G2D: Boosting Multimodal Learning with Gradient-Guided Distillation
Mohammed Rakib, Arunkumar Bagavathi
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 SMP (ablation) | G2D loss · CREMA-D | accuracy (%): 78.63 -> 85.89 | Improved |
| add SMP (ablation) | G2D loss · AV-MNIST | accuracy (%): 72.76 -> 73.03 | Improved |
| add SMP (ablation) | G2D loss · UR-Funny | accuracy (%): 63.78 -> 65.49 | Improved |
| add SMP (ablation) | Joint-Train · CREMA-D | accuracy (%): 67.47 -> 80.78 | Improved |
| add SMP (ablation) | UMT · CREMA-D | accuracy (%): 67.61 -> 82.39 | Improved |
| add G2D loss (feature+logit distillation) instead of UMT loss (ablation) | SMP training · CREMA-D | accuracy (%): 67.61 -> 78.63 | Improved |
| use complete suppression (SMP) instead of partial suppression (ablation) | G2D · CREMA-D | accuracy (%): 81.99 -> 85.89 | Improved |
| use complete suppression (SMP) instead of partial suppression (ablation) | G2D · VGGSound | accuracy (%): 51.16 -> 53.82 | Improved |
| use late fusion instead of FiLM (ablation) | G2D · VGGSound | accuracy (%): 48.11 -> 53.82 | Improved |
| use late fusion instead of Sum (ablation) | G2D · CREMA-D | accuracy (%): 81.59 -> 85.89 | Improved |
| use cross-attention instead of late fusion (ablation) | G2D · CREMA-D | accuracy (%): 85.89 -> 85.35 | Hurt |
| increase tau1 dedicated epochs for weak modality from 0 to 150 (ablation) | G2D · CREMA-D | accuracy (%): 78.63 -> 85.89 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- ReconBoost — ReconBoost: Boosting Can Achieve Modality Reconcilement (ICML 2024) · better on 1 of 1 shared result
- MMPareto — MMPareto: Boosting Multimodal Learning with Innocent Unimodal Assistance (ICML 2024) · better on 1 of 1 shared result
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 0.89)
- Rethinking Multimodal Learning from the Perspective of Mitigating Classification Ability Disproportion (NeurIPS 2025 · similarity 0.88)
- Modality-Aware SAM: Sharpness-Aware-Minimization Driven Gradient Modulation for Harmonized Multimodal Learning (NeurIPS 2025 · similarity 0.88)
- Asymmetric Reinforcing Against Multi-Modal Representation Bias (AAAI 2025 · similarity 0.87)
- Facilitating Multimodal Classification via Dynamically Learning Modality Gap (NeurIPS 2024 · similarity 0.87)
- Fusion or Confusion? Multimodal Complexity Is Not All You Need (NeurIPS 2026 · 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.
