Research map / Vision-language and multimodal / Human and embodied modeling
UViM: A Unified Modeling Approach for Vision with Learned Guiding Codes
Alexander Kolesnikov, André Susano Pinto, Lucas Beyer, Xiaohua Zhai, Jeremiah J. Harmsen, Neil Houlsby
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 |
|---|---|---|---|
| initialize LM encoder with pre-trained weights (ablation) | UViM · COCO Panoptic holdout | PQ: 41.4 -> 43.7 | Improved |
| scale LM from Base to Large with pre-training (ablation) | UViM · COCO Panoptic holdout | PQ: 41.9 -> 43.7 | Improved |
| scale LM from Base to Large when training from scratch (ablation) | UViM · COCO Panoptic holdout | PQ: 41.0 -> 41.4 | No clear effect |
| add code dropout in stage I (ablation) | UViM · COCO Panoptic holdout | PQ (final model): 42.2 -> 43.7 | Improved |
| add code dropout in stage I (ablation) | UViM · NYU Depth v2 | RMSE | No clear effect |
| add restricted oracle guiding code (ablation) | base model · COCO Panoptic holdout | PQ: 25.5 -> 75.7 | Improved |
| make LM autoregressive instead of single-pass (ablation) | UViM LM · COCO Panoptic holdout | PQ: 33.3 -> 43.7 | Improved |
| make LM autoregressive instead of single-pass (ablation) | UViM LM · NYU Depth v2 | RMSE: 0.55 -> 0.47 | Improved |
| feed image input to base model in stage I (ablation) | UViM · COCO Panoptic holdout | PQ: 39.1 -> 43.7 | Improved |
| feed image input to base model in stage I (ablation) | UViM · NYU Depth v2 | RMSE: within 0.005 RMSE | No clear effect |
| vary guiding code length and dictionary size (ablation) | UViM · COCO Panoptic holdout | PQ | Mixed |
Later papers that compared with this method
2 later accepted papers report UViM as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- UNIFIED-IO: A Unified Model for Vision, Language, and Multi-modal Tasks (ICLR 2023 · better on 1 of 1 shared result)
- All in Tokens: Unifying Output Space of Visual Tasks via Soft Token (ICCV 2023 · 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
- UVU: Improving Multimodal Understanding via Vision-Language Unified Autoregressive Paradigm (CVPR 2026 · similarity 0.84)
- UniBench: Visual Reasoning Requires Rethinking Vision-Language Beyond Scaling (NeurIPS 2024 · similarity 0.84)
- Explanatory Instructions: Towards Unified Vision Tasks Understanding and Zero-shot Generalization (ICML 2025 · similarity 0.84)
- Data-efficient Large Vision Models through Sequential Autoregression (ICML 2024 · similarity 0.84)
- Linear Differential Vision Transformer: Learning Visual Contrasts via Pairwise Differentials (NeurIPS 2025 · similarity 0.84)
- AutoVP: An Automated Visual Prompting Framework and Benchmark (ICLR 2024 · similarity 0.84)
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.
