Research map / Vision-language and multimodal / Cross-modal learning
Bridging Domain Generalization to Multimodal Domain Generalization via Unified Representations
Hai Huang, Yan Xia, Sashuai Zhou, Hanting Wang, Shulei Wang, Zhou Zhao
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 |
|---|---|---|---|
| apply Mixup in the unified representation (UR-Mixup) instead of directly | Mixup · EPIC-Kitchens (V+A, multi-source) | top-1 accuracy (%): 61.28 -> 64.77 | Improved |
| apply JiGen in the unified representation (UR-JiGen) | JiGen · EPIC-Kitchens (V+A, multi-source) | top-1 accuracy (%): 61.08 -> 63.63 | Improved |
| apply IBN in the unified representation (UR-IBN) | IBN-Net · EPIC-Kitchens (V+A, multi-source) | top-1 accuracy (%): 60.00 -> 61.41 | Improved |
| apply Mixup in the unified representation (UR-Mixup) | Mixup · HAC (V+A, multi-source) | top-1 accuracy (%): 66.41 -> 72.36 | Improved |
| apply Mixup in the unified representation (UR-Mixup) | Mixup · EPIC-Kitchens (V+A+F, multi-source) | top-1 accuracy (%): 63.42 -> 67.79 | Improved |
| apply Mixup in the unified representation (UR-Mixup) | Mixup · HAC (V+A+F, multi-source) | top-1 accuracy (%): 67.78 -> 73.40 | Improved |
| UR-Mixup versus CMRF in single-source DG | UR-Mixup · EPIC-Kitchens (V+A+F, single-source) | top-1 accuracy (%): 60.12 -> 60.84 | Improved |
| use supervised instead of unsupervised contrastive learning (ablation) | UR-Mixup · EPIC-Kitchens (V+A) | top-1 accuracy (%): 61.38 -> 62.23 | Improved |
| add CLUB decoupling to supervised contrastive learning instead of nothing (ablation) | UR-Mixup · EPIC-Kitchens (V+A) | top-1 accuracy (%): 62.23 -> 64.77 | Improved |
| use CLUB decoupling instead of MSE decoupling (ablation) | UR-Mixup with SCL · EPIC-Kitchens (V+A) | top-1 accuracy (%): 63.82 -> 64.77 | Improved |
| use random mixing ratio instead of fixed 0.5 (ablation) | UR-Mixup · EPIC-Kitchens (V+A) | top-1 accuracy (%): 63.48 -> 64.77 | Improved |
| use Jigsaw number 256 instead of 128 (ablation) | UR-JiGen · EPIC-Kitchens (V+A) | top-1 accuracy (%): 62.79 -> 63.63 | Improved |
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.
- SimMMDG — SimMMDG: A Simple and Effective Framework for Multi-modal Domain Generalization (NeurIPS 2023) · better on 3 of 3 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
- Cross-modal Representation Flattening for Multi-modal Domain Generalization (NeurIPS 2024 · similarity 0.93)
- Multimodal Unsupervised Domain Generalization by Retrieving Across the Modality Gap (ICLR 2025 · similarity 0.92)
- Are We Making Progress in Multimodal Domain Generalization? A Comprehensive Benchmark Study (NeurIPS 2026 · similarity 0.90)
- PracticalDG: Perturbation Distillation on Vision-Language Models for Hybrid Domain Generalization (CVPR 2024 · similarity 0.90)
- Reasoning-Driven Multimodal LLM for Domain Generalization (ICLR 2026 · similarity 0.89)
- Is Large-scale Pretraining the Secret to Good Domain Generalization? (ICLR 2025 · similarity 0.89)
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.
