Research map / Vision-language and multimodal / Human and embodied modeling
Cheers: Decoupling Patch Details from Semantic Representations Enables Unified Multimodal Comprehension and Generation
Yichen Zhang, Da Peng, Zonghao Guo, zijian zhang, Xuesong Yang, Tong Sun, Shichu Sun, Yidan Zhang, Yanghao Li, Haiyan Zhao, Wang Xu, Qi Shi and 10 more
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 high-frequency injection (HFI) (ablation) | Cheers · GenEval | Overall: 0.17 -> 0.30 | Improved |
| add high-frequency injection (HFI) (ablation) | Cheers · DPG-Bench | Overall: 39.11 -> 51.63 | Improved |
| add high-frequency injection (HFI) (ablation) | Cheers · MMBench | accuracy: 66.3 -> 67.1 | No clear effect |
| add high-frequency injection (HFI) (ablation) | Cheers · SEEDBench | accuracy: 70.0 -> 69.8 | No clear effect |
| add high-frequency injection (HFI) (ablation) | Cheers · POPE | accuracy: 86.2 -> 87.5 | No clear effect |
| add generation training jointly with understanding (ablation) | Cheers · MMBench | accuracy: 65.2 -> 67.1 | No clear effect |
| add generation training jointly with understanding (ablation) | Cheers · SEEDBench | accuracy: 70.8 -> 69.8 | No clear effect |
| add generation training jointly with understanding (ablation) | Cheers · ChartQA | accuracy: 58.5 -> 59.9 | No clear effect |
| reconstruct pixels with VAE decoder before semantic encoding (ablation) | Cheers · ChartQA (256x256) | accuracy: 14.2 -> 42.1 | Improved |
| reconstruct pixels with VAE decoder before semantic encoding (ablation) | Cheers · OCRBench (256x256) | score: 2.2 -> 31.5 | Improved |
| scale alignment data from 558K to 4.6M without pixel reconstruction (ablation) | Cheers · OCRBench (256x256) | score: 2.5 -> 2.2 | No clear effect |
| scale alignment data from 558K to 4.6M without pixel reconstruction (ablation) | Cheers · MMBench (256x256) | accuracy: 23.2 -> 42.8 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- Show-o2-1.5B — Show-o: One Single Transformer to Unify Multimodal Understanding and Generation (ICLR 2025) · better on 0 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
- Vision as a Dialect: Unifying Visual Understanding and Generation via Text-Aligned Representations (NeurIPS 2025 · similarity 0.92)
- WEAVE: Unleashing and Benchmarking the In-context Interleaved Comprehension and Generation (CVPR 2026 · similarity 0.90)
- The Narrow Gate: Localized Image-Text Communication in Native Multimodal Models (NeurIPS 2025 · similarity 0.89)
- Seeing What Matters: A Training-Free Self-Guided Framework for Multimodal Detail Perception and Reasoning (CVPR 2026 · similarity 0.89)
- MMIE: Massive Multimodal Interleaved Comprehension Benchmark for Large Vision-Language Models (ICLR 2025 · similarity 0.89)
- TokenFlow: Unified Image Tokenizer for Multimodal Understanding and Generation (CVPR 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.
