Research map / Vision-language and multimodal / Multimodal LLMs
UniReflect: Self-Reflection Tuning for Unified Multimodal Understanding and Generation
Cong Wei, Zepeng Huang, Haoxian Tan, liang shuang, Lishuai Gao, Pengfei Yan, Xiaoming Wei
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 self-reflection test-time scaling | UniReflect · GenEval | overall: 0.83 -> 0.89 | Improved |
| add self-reflection test-time scaling | UniReflect · GEdit-Bench-EN | overall: 5.14 -> 6.07 | Improved |
| add self-reflection TTS | UniReflect · GenEval++ | overall: 0.654 -> 0.680 | Improved |
| use in-model verifier instead of OmniVerifier for TTS (ablation) | UniReflect-TTS · GenEval | overall: 0.86 -> 0.89 | Improved |
| add Semantic Perceiver (ablation) | UniReflect · GenEval | overall: 0.77 -> 0.81 | Improved |
| add decoupling verification (ablation) | UniReflect · ViVerBench | overall: 0.525 -> 0.569 | Improved |
| add decoupling verification (ablation) | UniReflect · GenEval | overall: 0.77 -> 0.78 | No clear effect |
| combine Semantic Perceiver and decoupling (ablation) | UniReflect · GenEval / ViVerBench | overall: 0.77 -> 0.83; 0.525 -> 0.579 | Improved |
| include verification task in training mix | UniReflect · GEdit-Bench-EN | overall: 5.14 vs 6.41-6.70 for baselines | Hurt |
| fine-tune 4B model on synthetic verification data vs OmniVerifier RL 7B | UniReflect · ViVerBench | overall: 0.559 -> 0.579 | 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.
- OmniGen2 — OmniGen2: Towards Instruction-Aligned Multimodal Generation (CVPR 2026) · better on 3 of 4 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
- WEAVE: Unleashing and Benchmarking the In-context Interleaved Comprehension and Generation (CVPR 2026 · similarity 0.90)
- MME-Unify: A Comprehensive Benchmark for Unified Multimodal Understanding and Generation Models (ICLR 2026 · similarity 0.90)
- DuoGen: Towards Autonomous Interleaved Multimodal Generation (CVPR 2026 · similarity 0.89)
- Learning to Generate via Understanding: Understanding-Driven Intrinsic Rewarding for Unified Multimodal Models (CVPR 2026 · similarity 0.89)
- ROVER: Benchmarking Reciprocal Cross-Modal Reasoning for Omnimodal Generation (ICLR 2026 · similarity 0.88)
- Vision as a Dialect: Unifying Visual Understanding and Generation via Text-Aligned Representations (NeurIPS 2025 · 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.
