---
title: "REVEALER — paper by Shi et al. (2026) — summary and ideas"
description: "Checking whether a text-to-image (T2I) output matches its prompt is usually done with single-number… 45 references, cited by 1 paper here."
authors: ["FuLin Shi", "Wenyi Xiao", "Leilei Gan", "Liang Ding", "Binchen"]
year: 2026
venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/2026.acl-long.2200/"
doi: "10.18653/v1/2026.acl-long.2200"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/shi-2026-revealer-reinforcement-guided-visual-reasoning/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/shi-2026-revealer-reinforcement-guided-visual-reasoning/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/16/) / REVEALER: Reinforcement-Guided Visual Reasoning for Element-Level Text-Image Alignment Evaluation

# REVEALER: Reinforcement-Guided Visual Reasoning for Element-Level Text-Image Alignment Evaluation

**Authors:** FuLin Shi, Wenyi Xiao, [Leilei Gan](https://md.research.elamur.ai/free/en/authors/leilei-gan/), [Liang Ding](https://md.research.elamur.ai/free/en/authors/liang-ding/) and Binchen · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.acl-long.2200/) · [PDF](https://aclanthology.org/2026.acl-long.2200.pdf) · [DOI 10.18653/v1/2026.acl-long.2200](https://doi.org/10.18653/v1/2026.acl-long.2200) · [OpenAlex](https://openalex.org/W7166824763)

**Citations:** 0 citations (OpenAlex, 2026-10-14) · 1 in our library, of the 149,439 papers whose bibliographies we read

## Summary
Checking whether a text-to-image (T2I) output matches its prompt is usually done with single-number similarity scores or template-driven QA pipelines. These give little insight into which specific prompt elements succeeded or failed. REVEALER addresses this by training an MLLM to assess each prompt element in three explicit steps. First it predicts bounding boxes for the element, then it writes a free-text rationale, and finally it outputs a score in [0, 1]. The authors built training trajectories automatically from the EvalMuse-40K training split, using Grounding DINO for localization and GPT-4o for reasoning text. The model, based on Qwen3-VL (4B and 8B), is first fine-tuned on these trajectories and then refined with GRPO. The GRPO reward combines output format, box accuracy and correctness of the alignment judgment. Across four benchmarks the system reaches state-of-the-art accuracy and outperforms Gemini 3 Pro on EvalMuse-40K. The approach makes T2I evaluation both more accurate and easier to inspect, since each verdict is tied to visible image evidence.

## Key points
- **Gains over a training-based baseline:** EvalMuse-40K improves by +13.3% and RichHF by +9.8%, while MHaluBench and GenAI-Bench each improve by +7.1%.
- **Comparison with a proprietary model and ablations:** REVEALER beats Gemini 3 Pro by +4.2% on EvalMuse-40K, and the ablations show a cumulative +19.6% over the base model.
- **Grounding evaluation set (D_BoxEval):** 2,000 manually box-annotated pairs drawn from EvalMuse-40K and RichHF.
  - Grounding DINO reaches 0.76 mIoU on this set.
  - GRPO training raises model mIoU by +0.11 (4B) and +0.15 (8B).
  - Grounding quality correlates strongly with alignment accuracy (Pearson r = 0.8773).
- **Strict Grounding Strategy:** raising the Grounding DINO confidence threshold from 0.35 to 0.55 increases empty-box outputs for abstract elements from 12.4% to 53.0%, giving +4.2% accuracy on abstract attributes.
- **Error propagation:** misleading boxes (mIoU \< 0.5) push reasoning hallucination to 46.2%, whereas falling back to global reasoning cuts it to 14.7%. Stated limitations include holistic qualities such as style or mood, and video.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Alibaba Group](https://md.research.elamur.ai/free/en/companies/alibaba-group/) — Company, Hangzhou, China · 168 works · 15,386 citations
- [Zhejiang University](https://md.research.elamur.ai/free/en/universities/zhejiang-university/) — University, Hangzhou, China · 172 works · 12,988 citations

📚 Shelves (2): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 2,869 works

🧭 Topics (1): [🧮 Reasoning](https://md.research.elamur.ai/free/en/topics/reasoning/) · 1,426 works

## Cite

Cite as: FuLin Shi, Wenyi Xiao, Leilei Gan, Liang Ding and Binchen. 2026. REVEALER: Reinforcement-Guided Visual Reasoning for Element-Level Text-Image Alignment Evaluation. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{shi2026revealer,
  title = {{REVEALER: Reinforcement-Guided Visual Reasoning for Element-Level Text-Image Alignment Evaluation}},
  author = {FuLin Shi, Wenyi Xiao, Leilei Gan, Liang Ding and Binchen},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.2200},
  url = {https://doi.org/10.18653/v1/2026.acl-long.2200}
}
```

- [← previous RETUYT-INCO at BEA 2026 Shared Task 2: Meta-prompting in Rubric-based Scoring for German](https://md.research.elamur.ai/free/en/papers/sastre-2026-retuyt-inco-bea-2026-shared/)
- [next → Revealing Procedural Reasoning Structures in Chain-of-Thought Training via Span-Level Gradient Organization](https://md.research.elamur.ai/free/en/papers/liu-2026-revealing-procedural-reasoning-structures-chain/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/16/)

## References

45 references, 15 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Jinze Bai, Shuai Bai, Shusheng Yang, Shijie Wang, Sinan Tan, Peng Wang, Junyang Lin, Chang Zhou, and Jingren Zhou. 2023. [Qwen-vl: A versatile vision-language model for understanding, localization, text reading, and beyond](https://md.research.elamur.ai/free/en/papers/bai-2023-qwen-vl-versatile-vision-language/). arXiv preprint arXiv:2308.12966.
2. Sule Bai, Mingxing Li, Yong Liu, Jing Tang, Haoji Zhang, Lei Sun, Xiangxiang Chu, and Yansong Tang. 2025. Univg-r1: Reasoning guided universal visual grounding with reinforcement learning. arXiv preprint arXiv:2505.14231. — [arXiv:2505.14231](https://arxiv.org/abs/2505.14231) [find](https://scholar.google.com/scholar?q=Sule+Bai%2C+Mingxing+Li%2C+Yong+Liu%2C+Jing+Tang%2C+Haoji+Zhang%2C+Lei+Sun%2C+Xiangxiang+Chu%2C+and+Yansong+Tang%2E+2025%2E+Univg%2Dr1%3A+Reasoning+guided+universal+visual+grounding+with+reinforcement+learning%2E+arXiv+prepr)
3. Xiang Chen, Chenxi Wang, Yida Xue, Ningyu Zhang, Xiaoyan Yang, Qiang Li, Yue Shen, Lei Liang, Jinjie Gu, and Huajun Chen. 2024. Unified hallucination detection for multimodal large language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1:… [find](https://scholar.google.com/scholar?q=Xiang+Chen%2C+Chenxi+Wang%2C+Yida+Xue%2C+Ningyu+Zhang%2C+Xiaoyan+Yang%2C+Qiang+Li%2C+Yue+Shen%2C+Lei+Liang%2C+Jinjie+Gu%2C+and+Huajun+Chen%2E+2024%2E+Unified+hallucination+detection+for+multimodal+large+language+models%2E+In)
4. Ethan Chern, Zhulin Hu, Steffi Chern, Siqi Kou, Jiadi Su, Yan Ma, Zhijie Deng, and Pengfei Liu. 2025. Thinking with generated images. arXiv preprint arXiv:2505.22525. — [arXiv:2505.22525](https://arxiv.org/abs/2505.22525) [find](https://scholar.google.com/scholar?q=Ethan+Chern%2C+Zhulin+Hu%2C+Steffi+Chern%2C+Siqi+Kou%2C+Jiadi+Su%2C+Yan+Ma%2C+Zhijie+Deng%2C+and+Pengfei+Liu%2E+2025%2E+Thinking+with+generated+images%2E+arXiv+preprint+arXiv%3A2505%2E22525%2E+%E2%80%94+%5BarXiv%3A2505%2E22525%5D%28https%3A%2F%2Farxi)
5. Jaemin Cho, Yushi Hu, Roopal Garg, Peter Anderson, Ranjay Krishna, Jason Baldridge, Mohit Bansal, Jordi Pont-Tuset, and Su Wang. 2024. Davidsonian scene graph: Improving reliability in fine-grained evaluation for text-to-image generation. arXiv preprint arXiv:2310.18235. [find](https://scholar.google.com/scholar?q=Jaemin+Cho%2C+Yushi+Hu%2C+Roopal+Garg%2C+Peter+Anderson%2C+Ranjay+Krishna%2C+Jason+Baldridge%2C+Mohit+Bansal%2C+Jordi+Pont%2DTuset%2C+and+Su+Wang%2E+2024%2E+Davidsonian+scene+graph%3A+Improving+reliability+in+fine%2Dgrained+ev)
6. Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, and Steven Hoi. 2023. [Instructblip: Towards general-purpose vision-language models with instruction tuning](https://md.research.elamur.ai/free/en/papers/dai-2023-instructblip-towards-general-purpose-vision/). Advances in neural information processing systems, 36.
7. Patrick Esser, Sumith Kulal, Andreas Blattmann, Rahim Entezari, Jonas Müller, Harry Saini, Yam Levi, Dominik Lorenz, Axel Sauer, Frederic Boesel, Dustin Podell, Tim Dockhorn, Zion English, Kyle Lacey, Alex Goodwin, Yannik Marek, and Robin Rombach. 2024. Scaling rectified flow transformers for high-r… — [arXiv:2403.03206](https://arxiv.org/abs/2403.03206) [find](https://scholar.google.com/scholar?q=Patrick+Esser%2C+Sumith+Kulal%2C+Andreas+Blattmann%2C+Rahim+Entezari%2C+Jonas+M%C3%BCller%2C+Harry+Saini%2C+Yam+Levi%2C+Dominik+Lorenz%2C+Axel+Sauer%2C+Frederic+Boesel%2C+Dustin+Podell%2C+Tim+Dockhorn%2C+Zion+English%2C+Kyle+Lacey%2C)
8. Shuhao Han, Haotian Fan, Jiachen Fu, Liang Li, Tao Li, Junhui Cui, Yunqiu Wang, Yang Tai, Jingwei Sun, Chun-Le Guo, and Chongyi Li. 2026. Evalmuse-40k: A fine-grained benchmark with comprehensive human annotations for text-to-image generation model alignment evaluation. In Proceedings of the AAAI Co… [find](https://scholar.google.com/scholar?q=Shuhao+Han%2C+Haotian+Fan%2C+Jiachen+Fu%2C+Liang+Li%2C+Tao+Li%2C+Junhui+Cui%2C+Yunqiu+Wang%2C+Yang+Tai%2C+Jingwei+Sun%2C+Chun%2DLe+Guo%2C+and+Chongyi+Li%2E+2026%2E+Evalmuse%2D40k%3A+A+fine%2Dgrained+benchmark+with+comprehensive+huma)
9. Jack Hessel, Ari Holtzman, Maxwell Forbes, Ronan Le Bras, and Yejin Choi. 2021. [CLIPScore: A reference-free evaluation metric for image captioning](https://md.research.elamur.ai/free/en/papers/hessel-2021-clipscore-reference-free-evaluation-metric/). In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 7514–7528, Online and Punta Cana, Dominican Republic. A…
10. Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. 2017. Gans trained by a two time-scale update rule converge to a local nash equilibrium. Advances in neural information processing systems, 30. [find](https://scholar.google.com/scholar?q=Martin+Heusel%2C+Hubert+Ramsauer%2C+Thomas+Unterthiner%2C+Bernhard+Nessler%2C+and+Sepp+Hochreiter%2E+2017%2E+Gans+trained+by+a+two+time%2Dscale+update+rule+converge+to+a+local+nash+equilibrium%2E+Advances+in+neural+i)
11. Seyed Mohammad Hadi Hosseini, Amir Mohammad Izadi, Ali Abdollahi, Armin Saghafian, and Mahdieh Soleymani Baghshah. 2025. T2i-fineeval: Fine-grained compositional metric for text-to-image evaluation. arXiv preprint arXiv:2503.11481. — [arXiv:2503.11481](https://arxiv.org/abs/2503.11481) [find](https://scholar.google.com/scholar?q=Seyed+Mohammad+Hadi+Hosseini%2C+Amir+Mohammad+Izadi%2C+Ali+Abdollahi%2C+Armin+Saghafian%2C+and+Mahdieh+Soleymani+Baghshah%2E+2025%2E+T2i%2Dfineeval%3A+Fine%2Dgrained+compositional+metric+for+text%2Dto%2Dimage+evaluation%2E+a)
12. Yushi Hu, Benlin Liu, Jungo Kasai, Yizhong Wang, Mari Ostendorf, Ranjay Krishna, and Noah A. Smith. 2023. Tifa: Accurate and interpretable text-to-image faithfulness evaluation with question answering. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 20406– 20… [find](https://scholar.google.com/scholar?q=Yushi+Hu%2C+Benlin+Liu%2C+Jungo+Kasai%2C+Yizhong+Wang%2C+Mari+Ostendorf%2C+Ranjay+Krishna%2C+and+Noah+A%2E+Smith%2E+2023%2E+Tifa%3A+Accurate+and+interpretable+text%2Dto%2Dimage+faithfulness+evaluation+with+question+answering)
13. Hang Hua, Ziyun Zeng, Yizhi Song, Yunlong Tang, Liu He, Daniel Aliaga, Wei Xiong, and Jiebo Luo. 2025. Mmig-bench: Towards comprehensive and explainable evaluation of multi-modal image generation models. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Hang+Hua%2C+Ziyun+Zeng%2C+Yizhi+Song%2C+Yunlong+Tang%2C+Liu+He%2C+Daniel+Aliaga%2C+Wei+Xiong%2C+and+Jiebo+Luo%2E+2025%2E+Mmig%2Dbench%3A+Towards+comprehensive+and+explainable+evaluation+of+multi%2Dmodal+image+generation+mode)
14. Kaiyi Huang, Chengqi Duan, Kaiyue Sun, Enze Xie, Zhenguo Li, and Xihui Liu. 2025a. T2i-compbench++: An enhanced and comprehensive benchmark for compositional text-to-image generation. arXiv preprint arXiv:2307.06350. [find](https://scholar.google.com/scholar?q=Kaiyi+Huang%2C+Chengqi+Duan%2C+Kaiyue+Sun%2C+Enze+Xie%2C+Zhenguo+Li%2C+and+Xihui+Liu%2E+2025a%2E+T2i%2Dcompbench%2B%2B%3A+An+enhanced+and+comprehensive+benchmark+for+compositional+text%2Dto%2Dimage+generation%2E+arXiv+preprint+a)
15. Ziwei Huang, Wanggui He, Quanyu Long, Yandi Wang, Haoyuan Li, Zhelun Yu, Fangxun Shu, Weilong Dai, Hao Jiang, Fei Wu, and Leilei Gan. 2025b. [T2I-FactualBench: Benchmarking the factuality of text-to-image models with knowledge-intensive concepts](https://md.research.elamur.ai/free/en/papers/huang-2025-t2i-factualbench-benchmarking-factuality-text/). In Proceedings of the 63rd Annual Meeting of the Assoc…
16. Yuval Kirstain, Adam Polyak, Uriel Singer, Shahbuland Matiana, Joe Penna, and Omer Levy. 2023. Pick-a-pic: An open dataset of user preferences for text-to-image generation. Advances in neural information processing systems, 36. [find](https://scholar.google.com/scholar?q=Yuval+Kirstain%2C+Adam+Polyak%2C+Uriel+Singer%2C+Shahbuland+Matiana%2C+Joe+Penna%2C+and+Omer+Levy%2E+2023%2E+Pick%2Da%2Dpic%3A+An+open+dataset+of+user+preferences+for+text%2Dto%2Dimage+generation%2E+Advances+in+neural+informat)
17. Max Ku, Dongfu Jiang, Cong Wei, Xiang Yue, and Wenhu Chen. 2024. [VIEScore: Towards explainable metrics for conditional image synthesis evaluation](https://md.research.elamur.ai/free/en/papers/ku-2024-viescore-explainable-metrics-conditional-image/). In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 12268–12290, Bangkok, Thailand.…
18. Baiqi Li, Zhiqiu Lin, Deepak Pathak, Jiayao Li, Yixin Fei, Kewen Wu, Tiffany Ling, Xide Xia, Pengchuan Zhang, Graham Neubig, and Deva Ramanan. 2024. Genai-bench: Evaluating and improving compositional text-to-visual generation. arXiv preprint arXiv:2406.13743. [find](https://scholar.google.com/scholar?q=Baiqi+Li%2C+Zhiqiu+Lin%2C+Deepak+Pathak%2C+Jiayao+Li%2C+Yixin+Fei%2C+Kewen+Wu%2C+Tiffany+Ling%2C+Xide+Xia%2C+Pengchuan+Zhang%2C+Graham+Neubig%2C+and+Deva+Ramanan%2E+2024%2E+Genai%2Dbench%3A+Evaluating+and+improving+compositional)
19. Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. 2023. [Blip-2: bootstrapping language-image pre-training with frozen image encoders and large language models](https://md.research.elamur.ai/free/en/papers/li-2023-blip-2-bootstrapping-language-image/). In Proceedings of the 40th International Conference on Machine Learning, ICML’23. JMLR.org.
20. Weiqi Li, Xuanyu Zhang, Shijie Zhao, Yabin Zhang, Junlin Li, Li Zhang, and Jian Zhang. 2025. Q-insight: Understanding image quality via visual reinforcement learning. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Weiqi+Li%2C+Xuanyu+Zhang%2C+Shijie+Zhao%2C+Yabin+Zhang%2C+Junlin+Li%2C+Li+Zhang%2C+and+Jian+Zhang%2E+2025%2E+Q%2Dinsight%3A+Understanding+image+quality+via+visual+reinforcement+learning%2E+In+The+Thirty%2Dninth+Annual+Confer)
21. Youwei Liang, Junfeng He, Gang Li, Peizhao Li, Arseniy Klimovskiy, Nicholas Carolan, Jiao Sun, Jordi Pont-Tuset, Sarah Young, Feng Yang, Junjie Ke, Krishnamurthy Dj Dvijotham, Katherine M. Collins, Yiwen Luo, Yang Li, Kai J Kohlhoff, Deepak Ramachandran, and Vidhya Navalpakkam. 2024. Rich human feed… [find](https://scholar.google.com/scholar?q=Youwei+Liang%2C+Junfeng+He%2C+Gang+Li%2C+Peizhao+Li%2C+Arseniy+Klimovskiy%2C+Nicholas+Carolan%2C+Jiao+Sun%2C+Jordi+Pont%2DTuset%2C+Sarah+Young%2C+Feng+Yang%2C+Junjie+Ke%2C+Krishnamurthy+Dj+Dvijotham%2C+Katherine+M%2E+Collins%2C+Yi)
22. Zhiqiu Lin, Deepak Pathak, Baiqi Li, Jiayao Li, Xide Xia, Graham Neubig, Pengchuan Zhang, and Deva Ramanan. 2024. Evaluating text-to-visual generation with image-to-text generation. arXiv preprint arXiv:2404.01291. [find](https://scholar.google.com/scholar?q=Zhiqiu+Lin%2C+Deepak+Pathak%2C+Baiqi+Li%2C+Jiayao+Li%2C+Xide+Xia%2C+Graham+Neubig%2C+Pengchuan+Zhang%2C+and+Deva+Ramanan%2E+2024%2E+Evaluating+text%2Dto%2Dvisual+generation+with+image%2Dto%2Dtext+generation%2E+arXiv+preprint+arX)
23. Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. 2023. [Visual instruction tuning](https://md.research.elamur.ai/free/en/papers/liu-2023-visual-instruction-tuning/). arXiv preprint arXiv:2304.08485.
24. Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Qing Jiang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, and Lei Zhang. 2025. [Grounding dino: Marrying dino with grounded pre-training for open-set object detection](https://md.research.elamur.ai/free/en/papers/liu-2023-grounding-dino-marrying-dino-grounded/). In Computer Vision – ECCV 2024, pages 38–55, Cham. Springer Natu…
25. Yujie Lu, Xianjun Yang, Xiujun Li, Xin Eric Wang, and William Yang Wang. 2023. Llmscore: Unveiling the power of large language models in text-to-image synthesis evaluation. Advances in neural information processing systems, 36. [find](https://scholar.google.com/scholar?q=Yujie+Lu%2C+Xianjun+Yang%2C+Xiujun+Li%2C+Xin+Eric+Wang%2C+and+William+Yang+Wang%2E+2023%2E+Llmscore%3A+Unveiling+the+power+of+large+language+models+in+text%2Dto%2Dimage+synthesis+evaluation%2E+Advances+in+neural+informat)
26. Zi-Ao Ma, Tian Lan, Rong-Cheng Tu, Shu-Hang Liu, Heyan Huang, Zhijing Wu, Chen Xu, and Xian-Ling Mao. 2025. T2i-eval-r1: Reinforcement learning-driven reasoning for interpretable text-to-image evaluation. arXiv preprint arXiv:2505.17897. — [arXiv:2505.17897](https://arxiv.org/abs/2505.17897) [find](https://scholar.google.com/scholar?q=Zi%2DAo+Ma%2C+Tian+Lan%2C+Rong%2DCheng+Tu%2C+Shu%2DHang+Liu%2C+Heyan+Huang%2C+Zhijing+Wu%2C+Chen+Xu%2C+and+Xian%2DLing+Mao%2E+2025%2E+T2i%2Deval%2Dr1%3A+Reinforcement+learning%2Ddriven+reasoning+for+interpretable+text%2Dto%2Dimage+evaluat)
27. Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, and 1 others. 2022. [Training language models to follow instructions with human feedback](https://md.research.elamur.ai/free/en/papers/ouyang-2022-training-language-models-follow-instructions/). Advances in neural information processing systems, 35.
28. Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas Müller, Joe Penna, and Robin Rombach. 2024. [Sdxl: Improving latent diffusion models for high-resolution image synthesis](https://md.research.elamur.ai/free/en/papers/podell-2023-sdxl-improving-latent-diffusion-models/). In International Conference on Learning Representations, volume 2024, pages 1862–1874.
29. Aditya Ramesh, Mikhail Pavlov, Gabriel Goh, Scott Gray, Chelsea Voss, Alec Radford, Mark Chen, and Ilya Sutskever. 2021. [Zero-shot text-to-image generation](https://md.research.elamur.ai/free/en/papers/ramesh-2021-zero-shot-text-image-generation/). In Proceedings of the 38th International Conference on Machine Learning, volume 139 of Proceedings of Machine Learning Research, pages 8821–883…
30. Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. 2022. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10684–10695. [find](https://scholar.google.com/scholar?q=Robin+Rombach%2C+Andreas+Blattmann%2C+Dominik+Lorenz%2C+Patrick+Esser%2C+and+Bj%C3%B6rn+Ommer%2E+2022%2E+High%2Dresolution+image+synthesis+with+latent+diffusion+models%2E+In+Proceedings+of+the+IEEE%2FCVF+Conference+on+Compu)
31. Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily Denton, Seyed Kamyar Seyed Ghasemipour, Burcu Karagol Ayan, S. Sara Mahdavi, Rapha Gontijo Lopes, Tim Salimans, Jonathan Ho, David J Fleet, and Mohammad Norouzi. 2022. [Photorealistic text-to-image diffusion models with deep lan…](https://md.research.elamur.ai/free/en/papers/saharia-2022-photorealistic-text-image-diffusion-models/)
32. Tim Salimans, Ian Goodfellow, Wojciech Zaremba, Vicki Cheung, Alec Radford, and Xi Chen. 2016. [Improved techniques for training gans](https://md.research.elamur.ai/free/en/papers/salimans-2016-improved-techniques-training-gans/). Advances in neural information processing systems, 29.
33. Zhihong Shao, Peiyi Wang, Qihao Zhu, Runxin Xu, Junxiao Song, Xiao Bi, Haowei Zhang, Mingchuan Zhang, Y. K. Li, Y. Wu, and Daya Guo. 2024. [A comprehensive survey of direct preference optimization: Datasets, theories, variants, and applications](https://md.research.elamur.ai/free/en/papers/shao-2024-deepseekmath-pushing-limits-mathematical-reasoning/). arXiv preprint arXiv:2402.03300.
34. Zhaochen Su, Linjie Li, Mingyang Song, Yunzhuo Hao, Zhengyuan Yang, Jun Zhang, Guanjie Chen, Jiawei Gu, Juntao Li, Xiaoye Qu, and Yu Cheng. 2025. Openthinkimg: Learning to think with images via visual tool reinforcement learning. arXiv preprint arXiv:2505.08617. — [arXiv:2505.08617](https://arxiv.org/abs/2505.08617) [find](https://scholar.google.com/scholar?q=Zhaochen+Su%2C+Linjie+Li%2C+Mingyang+Song%2C+Yunzhuo+Hao%2C+Zhengyuan+Yang%2C+Jun+Zhang%2C+Guanjie+Chen%2C+Jiawei+Gu%2C+Juntao+Li%2C+Xiaoye+Qu%2C+and+Yu+Cheng%2E+2025%2E+Openthinkimg%3A+Learning+to+think+with+images+via+visual)
35. Zhiyu Tan, Xiaomeng Yang, Luozheng Qin, Mengping Yang, Cheng Zhang, and Hao Li. 2024. Evalalign: Supervised fine-tuning multimodal llms with human-aligned data for evaluating text-to-image models. arXiv preprint arXiv:2406.16562. the OpenAI Team. 2024. [Gpt-4 technical report](https://md.research.elamur.ai/free/en/papers/openai-2023-gpt-4-technical-report/). arXiv preprint arXiv:23…
36. Rong-Cheng Tu, Zi-Ao Ma, Tian Lan, Yuehao Zhao, Heyan Huang, and Xian-Ling Mao. 2025. Automatic evaluation for text-to-image generation: Task-decomposed framework, distilled training, and meta-evaluation benchmark. In Proceedings of the 63rd Annual Meeting of the Association for Computational Lingui… [find](https://scholar.google.com/scholar?q=Rong%2DCheng+Tu%2C+Zi%2DAo+Ma%2C+Tian+Lan%2C+Yuehao+Zhao%2C+Heyan+Huang%2C+and+Xian%2DLing+Mao%2E+2025%2E+Automatic+evaluation+for+text%2Dto%2Dimage+generation%3A+Task%2Ddecomposed+framework%2C+distilled+training%2C+and+meta%2Devaluat)
37. Yibin Wang, Zhimin Li, Yuhang Zang, Chunyu Wang, Qinglin Lu, Cheng Jin, and Jiaqi Wang. 2025a. Unified multimodal chain-of-thought reward model through reinforcement fine-tuning. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Yibin+Wang%2C+Zhimin+Li%2C+Yuhang+Zang%2C+Chunyu+Wang%2C+Qinglin+Lu%2C+Cheng+Jin%2C+and+Jiaqi+Wang%2E+2025a%2E+Unified+multimodal+chain%2Dof%2Dthought+reward+model+through+reinforcement+fine%2Dtuning%2E+In+The+Thirty%2Dninth+A)
38. Yibin Wang, Yuhang Zang, Hao Li, Cheng Jin, and Jiaqi Wang. 2025b. Unified reward model for multimodal understanding and generation. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Yibin+Wang%2C+Yuhang+Zang%2C+Hao+Li%2C+Cheng+Jin%2C+and+Jiaqi+Wang%2E+2025b%2E+Unified+reward+model+for+multimodal+understanding+and+generation%2E+In+The+Thirty%2Dninth+Annual+Conference+on+Neural+Information+Process)
39. Junfei Wu, Jian Guan, Kaituo Feng, Qiang Liu, Shu Wu, Liang Wang, Wei Wu, and Tieniu Tan. 2025. Reinforcing spatial reasoning in vision-language models with interwoven thinking and visual drawing. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Junfei+Wu%2C+Jian+Guan%2C+Kaituo+Feng%2C+Qiang+Liu%2C+Shu+Wu%2C+Liang+Wang%2C+Wei+Wu%2C+and+Tieniu+Tan%2E+2025%2E+Reinforcing+spatial+reasoning+in+vision%2Dlanguage+models+with+interwoven+thinking+and+visual+drawing%2E+In+)
40. Wenyi Xiao, Ziwei Huang, Leilei Gan, Wanggui He, Haoyuan Li, Zhelun Yu, Fangxun Shu, Hao Jiang, and Linchao Zhu. 2025a. Detecting and mitigating hallucination in large vision language models via fine-grained ai feedback. In Proceedings of the AAAI Conference on Artificial Intelligence. [find](https://scholar.google.com/scholar?q=Wenyi+Xiao%2C+Ziwei+Huang%2C+Leilei+Gan%2C+Wanggui+He%2C+Haoyuan+Li%2C+Zhelun+Yu%2C+Fangxun+Shu%2C+Hao+Jiang%2C+and+Linchao+Zhu%2E+2025a%2E+Detecting+and+mitigating+hallucination+in+large+vision+language+models+via+fine%2D)
41. Wenyi Xiao, Zechuan Wang, Leilei Gan, Shuai Zhao, Zongrui Li, Ruirui Lei, Wanggui He, Luu Anh Tuan, Long Chen, Hao Jiang, Zhou Zhao, and Fei Wu. 2025b. A comprehensive survey of direct preference optimization: Datasets, theories, variants, and applications. arXiv preprint arXiv:2410.15595. — [arXiv:2410.15595](https://arxiv.org/abs/2410.15595) [find](https://scholar.google.com/scholar?q=Wenyi+Xiao%2C+Zechuan+Wang%2C+Leilei+Gan%2C+Shuai+Zhao%2C+Zongrui+Li%2C+Ruirui+Lei%2C+Wanggui+He%2C+Luu+Anh+Tuan%2C+Long+Chen%2C+Hao+Jiang%2C+Zhou+Zhao%2C+and+Fei+Wu%2E+2025b%2E+A+comprehensive+survey+of+direct+preference+opti)
42. Michal Yarom, Yonatan Bitton, Soravit Changpinyo, Roee Aharoni, Jonathan Herzig, Oran Lang, Eran Ofek, and Idan Szpektor. 2023. What you see is what you read? improving text-image alignment evaluation. Advances in neural information processing systems, 36. [find](https://scholar.google.com/scholar?q=Michal+Yarom%2C+Yonatan+Bitton%2C+Soravit+Changpinyo%2C+Roee+Aharoni%2C+Jonathan+Herzig%2C+Oran+Lang%2C+Eran+Ofek%2C+and+Idan+Szpektor%2E+2023%2E+What+you+see+is+what+you+read%3F+improving+text%2Dimage+alignment+evaluation)
43. Yufei Zhan, Yousong Zhu, Shurong Zheng, Hongyin Zhao, Fan Yang, Ming Tang, and Jinqiao Wang. 2025. Vision-r1: Evolving human-free alignment in large vision-language models via vision-guided reinforcement learning. arXiv preprint arXiv:2503.18013. — [arXiv:2503.18013](https://arxiv.org/abs/2503.18013) [find](https://scholar.google.com/scholar?q=Yufei+Zhan%2C+Yousong+Zhu%2C+Shurong+Zheng%2C+Hongyin+Zhao%2C+Fan+Yang%2C+Ming+Tang%2C+and+Jinqiao+Wang%2E+2025%2E+Vision%2Dr1%3A+Evolving+human%2Dfree+alignment+in+large+vision%2Dlanguage+models+via+vision%2Dguided+reinforcem)
44. Xintong Zhang, Zhi Gao, Bofei Zhang, Pengxiang Li, Xiaowen Zhang, Yang Liu, Tao Yuan, Yuwei Wu, Yunde Jia, Song-Chun Zhu, and Qing Li. 2025. Chain-of-focus: Adaptive visual search and zooming for multimodal reasoning via rl. arXiv preprint arXiv:2505.15436. — [arXiv:2505.15436](https://arxiv.org/abs/2505.15436) [find](https://scholar.google.com/scholar?q=Xintong+Zhang%2C+Zhi+Gao%2C+Bofei+Zhang%2C+Pengxiang+Li%2C+Xiaowen+Zhang%2C+Yang+Liu%2C+Tao+Yuan%2C+Yuwei+Wu%2C+Yunde+Jia%2C+Song%2DChun+Zhu%2C+and+Qing+Li%2E+2025%2E+Chain%2Dof%2Dfocus%3A+Adaptive+visual+search+and+zooming+for+mult)
45. Ziwei Zheng, Michael Yang, Jack Hong, Chenxiao Zhao, Guohai Xu, Le Yang, Chao Shen, and Xing Yu. 2025. Deepeyes: Incentivizing "thinking with images" via reinforcement learning. arXiv preprint arXiv:2505.14362. — [arXiv:2505.14362](https://arxiv.org/abs/2505.14362) [find](https://scholar.google.com/scholar?q=Ziwei+Zheng%2C+Michael+Yang%2C+Jack+Hong%2C+Chenxiao+Zhao%2C+Guohai+Xu%2C+Le+Yang%2C+Chao+Shen%2C+and+Xing+Yu%2E+2025%2E+Deepeyes%3A+Incentivizing+%22thinking+with+images%22+via+reinforcement+learning%2E+arXiv+preprint+arXiv%3A2)

*Updated 31 October 2026*