Research map / Vision-language and multimodal / Human and embodied modeling
Exploring the Effect of Primitives for Compositional Generalization in Vision-and-Language
Chuanhao Li, Zhen Li, Chenchen Jing, Yunde Jia, Yuwei Wu
This paper tried a common change: Add a contrastive loss
Add contrastive loss to training with equivariant and invariant samples · Charades-CG Novel-Composition · R1@0.5: 45.55 -> 46.54 Improved
What 1,317 accepted papers report for this change
- 1,253 improved
- 22 no clear effect
- 16 mixed
- 26 hurt
- Beyond Looking Up, Try Looking Around: Harmonizing Global Structure and Local Consistency in Optimal Transport for Short Text Clustering ICML 2026 · add contrastive learning · SearchSnippets · ACC: 56.65 -> 80.12Improved
- SleepLM: Natural-Language Intelligence for Human Sleep ICML 2026 · add contrastive objective to captioning (CoCa vs Cap) · sleep event IoU · IoU: 29.2 -> 27.4Hurt
- LIMSSR: LLM-Driven Sequence-to-Score Reasoning under Training-Time Incomplete Multimodal Observations ICML 2026 · add contrastive loss L_con · FS1000 average · Spearman correlation: 0.759 -> 0.789Improved
See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.
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 the primitive-based framework | MS-2D-TAN · Charades-CG Novel-Composition | R1@0.5: 43.17 -> 46.54 | Improved |
| Add the primitive-based framework | MS-2D-TAN · Charades-CG Novel-Composition | mIoU: 38.06 -> 40.00 | Improved |
| Add the primitive-based framework | MS-2D-TAN · Charades-STA test | R1@0.7: 37.34 -> 38.49 | Improved |
| Add the primitive-based framework | GLT · CLOSURE | accuracy: 96.1 -> 98.4 | Improved |
| Add the primitive-based framework | GLT · CLEVR | accuracy: 99.1 -> 99.1 | No clear effect |
| Use equivariant samples only (ablation) | MS-2D-TAN · Charades-CG Novel-Composition | R1@0.5: 43.17 -> 45.38 | Improved |
| Use invariant samples only (ablation) | MS-2D-TAN · Charades-CG Novel-Composition | R1@0.5: 43.17 -> 43.90 | Improved |
| Combine equivariant and invariant samples (ablation) | MS-2D-TAN · Charades-CG Novel-Composition | R1@0.5: 43.17 -> 45.55 | Improved |
| Add contrastive loss to training with equivariant and invariant samples (ablation) | MS-2D-TAN · Charades-CG Novel-Composition | R1@0.5: 45.55 -> 46.54 | Improved |
Closest papers by meaning
- Does Data Scaling Lead to Visual Compositional Generalization? (ICML 2025 · similarity 0.90)
- Not Only Text: Exploring Compositionality of Visual Representations in Vision-Language Models (CVPR 2025 · similarity 0.89)
- When and How Does CLIP Enable Domain and Compositional Generalization? (ICML 2025 · similarity 0.89)
- Iterated Learning Improves Compositionality in Large Vision-Language Models (CVPR 2024 · similarity 0.89)
- When does compositional structure yield compositional generalization? A kernel theory. (ICLR 2025 · similarity 0.89)
- Investigating Compositional Challenges in Vision-Language Models for Visual Grounding (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.
