---
title: "CSMCIR — paper by Qian et al. (2026) — summary and ideas"
description: "The authors argue that current CIR systems handle queries (image plus text) and targets (image only) with different encoders. 58 references."
authors: ["Zhipeng Qian", "Zihan Liang", "Yufei Ma", "Ben Chen", "Huangyu Dai", "Yiwei Ma", "Jiayi Ji", "Chenyi Lei", "Han Li", "Xiaoshuai Sun"]
year: 2026
venue: "Findings of the Association for Computational Linguistics: ACL 2026"
source: "https://aclanthology.org/2026.findings-acl.216/"
doi: "10.18653/v1/2026.findings-acl.216"
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/qian-2026-csmcir-cot-enhanced-symmetric-alignment/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/qian-2026-csmcir-cot-enhanced-symmetric-alignment/)

[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/26/) / CSMCIR: CoT-Enhanced Symmetric Alignment with Memory Bank for Composed Image Retrieval

# CSMCIR: CoT-Enhanced Symmetric Alignment with Memory Bank for Composed Image Retrieval

**Authors:** Zhipeng Qian, [Zihan Liang](https://md.research.elamur.ai/free/en/authors/zihan-liang/), Yufei Ma, Ben Chen, Huangyu Dai, Yiwei Ma, Jiayi Ji, Chenyi Lei, [Han Li](https://md.research.elamur.ai/free/en/authors/han-li/) and Xiaoshuai Sun · **Year:** 2026 · **Venue:** [Findings of the Association for Computational Linguistics: ACL 2026](https://md.research.elamur.ai/free/en/venues/findings/) ([Findings of the Association for Computational Linguistics 2026](https://md.research.elamur.ai/free/en/venues/findings/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.findings-acl.216/) · [PDF](https://aclanthology.org/2026.findings-acl.216.pdf) · [DOI 10.18653/v1/2026.findings-acl.216](https://doi.org/10.18653/v1/2026.findings-acl.216) · [OpenAlex](https://openalex.org/W7166865510)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
Composed Image Retrieval (CIR) finds a target image from a query that pairs a reference image with a modification text. The authors argue that current CIR systems handle queries (image plus text) and targets (image only) with different encoders. This leaves the two sides in separate embedding regions, which training only partly reconciles. CSMCIR fixes this mismatch in three ways:

- An MLLM, guided by a Multi-level Chain-of-Thought (MCoT) prompt, writes discriminative captions for target images offline. Each target thus becomes an image-text pair, like the query.
- One shared-parameter Q-Former encodes both sides in a symmetric dual-tower model built on SPRC.
- An entropy-aware Memory Bank with temporal decay stores captions and image embeddings. It re-encodes them with the current model to produce consistent hard negatives.

On Fashion-IQ, CIRR, Shoes and LaSCO, the method reports state-of-the-art recall and better training efficiency. Because captions are precomputed, retrieval has no added inference latency, which matters for deployments such as e-commerce search.

## Key points
- With ViT-G, CSMCIR averages 67.17 on Fashion-IQ (R@10/R@50) and 82.58 on CIRR. With ViT-L it reaches 64.45 on Fashion-IQ, ahead of QURE (63.04) and SPRC (62.21).
- In ablation, the Entropy-Aware Memory Bank (EAMB) raised Fashion-IQ performance from 65.92 to 66.65. Removing Temporal Decay hurt results more than removing the entropy component.
- A naive MoCo-style FIFO memory bank degrades quickly once it exceeds the batch size. The proposed bank levels off at size 512.
- t-SNE plots show that CSMCIR's query and target embeddings are already closer before training and nearly merge after training.
- Stated limitations are the upfront cost of generating captions and evaluation limited to fashion and general-object domains. Code is available at github.com/qzp2018/CSMCIR.

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

📚 Shelves (3): [🖥️ 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 · [🗄️ Information Systems](https://md.research.elamur.ai/free/en/shelves/computer-science/information-systems/) · 1,437 works

🧭 Topics (1): [🔍 Information retrieval & search](https://md.research.elamur.ai/free/en/topics/information-retrieval/) · 473 works

## Cite

Cite as: Zhipeng Qian, Zihan Liang, Yufei Ma, Ben Chen, Huangyu Dai, Yiwei Ma, Jiayi Ji, Chenyi Lei, Han Li and Xiaoshuai Sun. 2026. CSMCIR: CoT-Enhanced Symmetric Alignment with Memory Bank for Composed Image Retrieval. Findings of the Association for Computational Linguistics: ACL 2026.

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

```bibtex
@inproceedings{qian2026csmcir,
  title = {{CSMCIR: CoT-Enhanced Symmetric Alignment with Memory Bank for Composed Image Retrieval}},
  author = {Zhipeng Qian, Zihan Liang, Yufei Ma, Ben Chen, Huangyu Dai, Yiwei Ma, Jiayi Ji, Chenyi Lei, Han Li and Xiaoshuai Sun},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.216},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.216}
}
```

- [← previous CSI: An Investigative Multi-Agent Framework for Explainable Short Video Fake News Detection](https://md.research.elamur.ai/free/en/papers/wang-2026-csi-investigative-multi-agent-framework/)
- [next → CSULoRA: Closest Safe Update Low-Rank Adaptation](https://md.research.elamur.ai/free/en/papers/marchenko-2026-csulora-closest-safe-update-low/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/26/)

## References

58 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Muhammad Umer Anwaar, Egor Labintcev, and Martin Kleinsteuber. 2021. Compositional learning of image-text query for image retrieval. In Proceedings of the IEEE/CVF Winter conference on Applications of Computer Vision, pages 1140–1149. [find](https://scholar.google.com/scholar?q=Muhammad+Umer+Anwaar%2C+Egor+Labintcev%2C+and+Martin+Kleinsteuber%2E+2021%2E+Compositional+learning+of+image%2Dtext+query+for+image+retrieval%2E+In+Proceedings+of+the+IEEE%2FCVF+Winter+conference+on+Applications+of)
2. 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/).
3. Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, and 1 others. 2025. Qwen2. 5-vl technical report. arXiv preprint arXiv:2502.13923. — [Qwen2.5-VL Technical Report](https://md.research.elamur.ai/free/en/papers/bai-2025-qwen2-5-vl-technical/)
4. Alberto Baldrati, Lorenzo Agnolucci, Marco Bertini, and Alberto Del Bimbo. 2023a. Zero-shot composed image retrieval with textual inversion. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 15338–15347. [find](https://scholar.google.com/scholar?q=Alberto+Baldrati%2C+Lorenzo+Agnolucci%2C+Marco+Bertini%2C+and+Alberto+Del+Bimbo%2E+2023a%2E+Zero%2Dshot+composed+image+retrieval+with+textual+inversion%2E+In+Proceedings+of+the+IEEE%2FCVF+International+Conference+on+)
5. Rinon Gal, Yuval Alaluf, Yuval Atzmon, Or Patashnik, Amit H Bermano, Gal Chechik, and Daniel Cohen-Or. 2022. An image is worth one word: Personalizing text-to-image generation using textual inversion. arXiv preprint arXiv:2208.01618. [find](https://scholar.google.com/scholar?q=Rinon+Gal%2C+Yuval+Alaluf%2C+Yuval+Atzmon%2C+Or+Patashnik%2C+Amit+H+Bermano%2C+Gal+Chechik%2C+and+Daniel+Cohen%2DOr%2E+2022%2E+An+image+is+worth+one+word%3A+Personalizing+text%2Dto%2Dimage+generation+using+textual+inversion%2E)
6. Kaiming He, Haoqi Fan, Yuxin Wu, Saining Xie, and Ross Girshick. 2020. [Momentum contrast for unsupervised visual representation learning](https://md.research.elamur.ai/free/en/papers/he-2019-momentum-contrast-unsupervised-visual-representation/). In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9729–9738.
7. Liqi He, Zuchao Li, Xiantao Cai, and Ping Wang. 2024. Multi-modal latent space learning for chain-of-thought reasoning in language models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 18180–18187. [find](https://scholar.google.com/scholar?q=Liqi+He%2C+Zuchao+Li%2C+Xiantao+Cai%2C+and+Ping+Wang%2E+2024%2E+Multi%2Dmodal+latent+space+learning+for+chain%2Dof%2Dthought+reasoning+in+language+models%2E+In+Proceedings+of+the+AAAI+Conference+on+Artificial+Intellige)
8. Young Kyun Jang, Donghyun Kim, Zihang Meng, Dat Huynh, and Ser-Nam Lim. 2024. Visual delta generator with large multi-modal models for semi-supervised composed image retrieval. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 16805–16814. [find](https://scholar.google.com/scholar?q=Young+Kyun+Jang%2C+Donghyun+Kim%2C+Zihang+Meng%2C+Dat+Huynh%2C+and+Ser%2DNam+Lim%2E+2024%2E+Visual+delta+generator+with+large+multi%2Dmodal+models+for+semi%2Dsupervised+composed+image+retrieval%2E+In+Proceedings+of+the+I)
9. Alberto Baldrati, Marco Bertini, Tiberio Uricchio, and Alberto Del Bimbo. 2022. Effective conditioned and composed image retrieval combining clip-based features. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 21466–21474. [find](https://scholar.google.com/scholar?q=Alberto+Baldrati%2C+Marco+Bertini%2C+Tiberio+Uricchio%2C+and+Alberto+Del+Bimbo%2E+2022%2E+Effective+conditioned+and+composed+image+retrieval+combining+clip%2Dbased+features%2E+In+Proceedings+of+the+IEEE%2FCVF+Confere)
10. Xintong Jiang, Yaxiong Wang, Mengjian Li, Yujiao Wu, Bingwen Hu, and Xueming Qian. 2024. Cala: Complementary association learning for augmenting composed image retrieval. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 2177–21… [find](https://scholar.google.com/scholar?q=Xintong+Jiang%2C+Yaxiong+Wang%2C+Mengjian+Li%2C+Yujiao+Wu%2C+Bingwen+Hu%2C+and+Xueming+Qian%2E+2024%2E+Cala%3A+Complementary+association+learning+for+augmenting+composed+image+retrieval%2E+In+Proceedings+of+the+47th+In)
11. Alberto Baldrati, Marco Bertini, Tiberio Uricchio, and Alberto Del Bimbo. 2023b. Composed image retrieval using contrastive learning and task-oriented clip-based features. ACM Transactions on Multimedia Computing, Communications and Applications, 20(3):1–24. [find](https://scholar.google.com/scholar?q=Alberto+Baldrati%2C+Marco+Bertini%2C+Tiberio+Uricchio%2C+and+Alberto+Del+Bimbo%2E+2023b%2E+Composed+image+retrieval+using+contrastive+learning+and+task%2Doriented+clip%2Dbased+features%2E+ACM+Transactions+on+Multimed)
12. Shyamgopal Karthik, Karsten Roth, Massimiliano Mancini, and Zeynep Akata. 2023. Vision-by-language for training-free compositional image retrieval. arXiv preprint arXiv:2310.09291. — [arXiv:2310.09291](https://arxiv.org/abs/2310.09291) [find](https://scholar.google.com/scholar?q=Shyamgopal+Karthik%2C+Karsten+Roth%2C+Massimiliano+Mancini%2C+and+Zeynep+Akata%2E+2023%2E+Vision%2Dby%2Dlanguage+for+training%2Dfree+compositional+image+retrieval%2E+arXiv+preprint+arXiv%3A2310%2E09291%2E+%E2%80%94+%5BarXiv%3A2310%2E09291)
13. Ben Chen, Siyuan Wang, Yufei Ma, Zihan Liang, Xuxin Zhang, Yue Lv, Ying Yang, Huangyu Dai, Lingtao Mao, Tong Zhao, and 1 others. 2026. Onesearch-v2: The latent reasoning enhanced self-distillation generative search framework. arXiv preprint arXiv:2603.24422. — [arXiv:2603.24422](https://arxiv.org/abs/2603.24422) [find](https://scholar.google.com/scholar?q=Ben+Chen%2C+Siyuan+Wang%2C+Yufei+Ma%2C+Zihan+Liang%2C+Xuxin+Zhang%2C+Yue+Lv%2C+Ying+Yang%2C+Huangyu+Dai%2C+Lingtao+Mao%2C+Tong+Zhao%2C+and+1+others%2E+2026%2E+Onesearch%2Dv2%3A+The+latent+reasoning+enhanced+self%2Ddistillation+gen)
14. Yanzhe Chen, Huasong Zhong, Xiangteng He, Yuxin Peng, Jiahuan Zhou, and Lele Cheng. 2024. Fash-ionern: enhance-and-refine network for composed fashion image retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 1228–1236. [find](https://scholar.google.com/scholar?q=Yanzhe+Chen%2C+Huasong+Zhong%2C+Xiangteng+He%2C+Yuxin+Peng%2C+Jiahuan+Zhou%2C+and+Lele+Cheng%2E+2024%2E+Fash%2Dionern%3A+enhance%2Dand%2Drefine+network+for+composed+fashion+image+retrieval%2E+In+Proceedings+of+the+AAAI+Confe)
15. Zhiwei Chen, Yupeng Hu, Zixu Li, Zhiheng Fu, Xuemeng Song, and Liqiang Nie. 2025. Offset: Segmentation-based focus shift revision for composed image retrieval. In Proceedings of the 33rd ACM International Conference on Multimedia, pages 6113–6122. [find](https://scholar.google.com/scholar?q=Zhiwei+Chen%2C+Yupeng+Hu%2C+Zixu+Li%2C+Zhiheng+Fu%2C+Xuemeng+Song%2C+and+Liqiang+Nie%2E+2025%2E+Offset%3A+Segmentation%2Dbased+focus+shift+revision+for+composed+image+retrieval%2E+In+Proceedings+of+the+33rd+ACM+Internati)
16. Zhangchi Feng, Richong Zhang, and Zhijie Nie. 2024. Improving composed image retrieval via contrastive learning with scaling positives and negatives. In Proceedings of the 32nd ACM International Conference on Multimedia, pages 1632–1641. [find](https://scholar.google.com/scholar?q=Zhangchi+Feng%2C+Richong+Zhang%2C+and+Zhijie+Nie%2E+2024%2E+Improving+composed+image+retrieval+via+contrastive+learning+with+scaling+positives+and+negatives%2E+In+Proceedings+of+the+32nd+ACM+International+Confe)
17. Sungyeon Kim, Xinliang Zhu, Xiaofan Lin, Muhammet Bastan, Douglas Gray, and Suha Kwak. 2025. Genius: A generative framework for universal multimodal search. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 19659– 19669. [find](https://scholar.google.com/scholar?q=Sungyeon+Kim%2C+Xinliang+Zhu%2C+Xiaofan+Lin%2C+Muhammet+Bastan%2C+Douglas+Gray%2C+and+Suha+Kwak%2E+2025%2E+Genius%3A+A+generative+framework+for+universal+multimodal+search%2E+In+Proceedings+of+the+Computer+Vision+and+P)
18. Jaehyun Kwak, Ramahdani Muhammad Izaaz Inhar, Se-Young Yun, and Sung-Ju Lee. 2025. Qure: Query-relevant retrieval through hard negative sampling in composed image retrieval. arXiv preprint arXiv:2507.12416. — [arXiv:2507.12416](https://arxiv.org/abs/2507.12416) [find](https://scholar.google.com/scholar?q=Jaehyun+Kwak%2C+Ramahdani+Muhammad+Izaaz+Inhar%2C+Se%2DYoung+Yun%2C+and+Sung%2DJu+Lee%2E+2025%2E+Qure%3A+Query%2Drelevant+retrieval+through+hard+negative+sampling+in+composed+image+retrieval%2E+arXiv+preprint+arXiv%3A2507%2E)
19. Matan Levy, Rami Ben-Ari, Nir Darshan, and Dani Lischinski. 2024. Data roaming and quality assessment for composed image retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 2991–2999. [find](https://scholar.google.com/scholar?q=Matan+Levy%2C+Rami+Ben%2DAri%2C+Nir+Darshan%2C+and+Dani+Lischinski%2E+2024%2E+Data+roaming+and+quality+assessment+for+composed+image+retrieval%2E+In+Proceedings+of+the+AAAI+Conference+on+Artificial+Intelligence%2C+vo)
20. 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 International conference on machine learning, pages 19730–19742. PMLR.
21. Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. 2022. [Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation](https://md.research.elamur.ai/free/en/papers/li-2022-blip-bootstrapping-language-image-pre/). In International conference on machine learning, pages 12888–12900. PMLR.
22. Shuxian Li, Changhao He, Xiting Liu, Joey Tianyi Zhou, Xi Peng, and Peng Hu. 2025a. Learning with noisy triplet correspondence for composed image retrieval. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 19628–19637. [find](https://scholar.google.com/scholar?q=Shuxian+Li%2C+Changhao+He%2C+Xiting+Liu%2C+Joey+Tianyi+Zhou%2C+Xi+Peng%2C+and+Peng+Hu%2E+2025a%2E+Learning+with+noisy+triplet+correspondence+for+composed+image+retrieval%2E+In+Proceedings+of+the+Computer+Vision+and+P)
23. Zixu Li, Zhiwei Chen, Haokun Wen, Zhiheng Fu, Yupeng Hu, and Weili Guan. 2025b. Encoder: Entity mining and modification relation binding for composed image retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 5101–5109. [find](https://scholar.google.com/scholar?q=Zixu+Li%2C+Zhiwei+Chen%2C+Haokun+Wen%2C+Zhiheng+Fu%2C+Yupeng+Hu%2C+and+Weili+Guan%2E+2025b%2E+Encoder%3A+Entity+mining+and+modification+relation+binding+for+composed+image+retrieval%2E+In+Proceedings+of+the+AAAI+Confer)
24. Zihan Liang, Yufei Ma, Zhipeng Qian, Huangyu Dai, Zihan Wang, Ben Chen, Chenyi Lei, Yuqing Ding, and Han Li. 2025. Uniecs: Unified multimodal ecommerce search framework with gated cross-modal fusion. In Proceedings of the 34th ACM International Conference on Information and Knowledge Management, CIK… [find](https://scholar.google.com/scholar?q=Zihan+Liang%2C+Yufei+Ma%2C+Zhipeng+Qian%2C+Huangyu+Dai%2C+Zihan+Wang%2C+Ben+Chen%2C+Chenyi+Lei%2C+Yuqing+Ding%2C+and+Han+Li%2E+2025%2E+Uniecs%3A+Unified+multimodal+ecommerce+search+framework+with+gated+cross%2Dmodal+fusion%2E+)
25. Weihuang Lin, Yiwei Ma, Jiayi Ji, Xiaoshuai Sun, and Rongrong Ji. 2025. Cir-cot: Towards interpretable composed image retrieval via end-to-end chain-of-thought reasoning. arXiv preprint arXiv:2510.08003. — [arXiv:2510.08003](https://arxiv.org/abs/2510.08003) [find](https://scholar.google.com/scholar?q=Weihuang+Lin%2C+Yiwei+Ma%2C+Jiayi+Ji%2C+Xiaoshuai+Sun%2C+and+Rongrong+Ji%2E+2025%2E+Cir%2Dcot%3A+Towards+interpretable+composed+image+retrieval+via+end%2Dto%2Dend+chain%2Dof%2Dthought+reasoning%2E+arXiv+preprint+arXiv%3A2510%2E080)
26. Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. 2023a. [Visual instruction tuning](https://md.research.elamur.ai/free/en/papers/liu-2023-visual-instruction-tuning/). Advances in neural information processing systems, 36:34892– 34916.
27. Zheyuan Liu, Weixuan Sun, Damien Teney, and Stephen Gould. 2023b. Candidate set re-ranking for composed image retrieval with dual multi-modal encoder. arXiv preprint arXiv:2305.16304. — [arXiv:2305.16304](https://arxiv.org/abs/2305.16304) [find](https://scholar.google.com/scholar?q=Zheyuan+Liu%2C+Weixuan+Sun%2C+Damien+Teney%2C+and+Stephen+Gould%2E+2023b%2E+Candidate+set+re%2Dranking+for+composed+image+retrieval+with+dual+multi%2Dmodal+encoder%2E+arXiv+preprint+arXiv%3A2305%2E16304%2E+%E2%80%94+%5BarXiv%3A2305%2E16)
28. Ilya Loshchilov and Frank Hutter. 2017. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101. [find](https://scholar.google.com/scholar?q=Ilya+Loshchilov+and+Frank+Hutter%2E+2017%2E+Decoupled+weight+decay+regularization%2E+arXiv+preprint+arXiv%3A1711%2E05101%2E)
29. Jiasen Lu, Dhruv Batra, Devi Parikh, and Stefan Lee. 2019. [Vilbert: Pretraining task-agnostic visiolinguistic representations for vision-and-language tasks](https://md.research.elamur.ai/free/en/papers/lu-2019-vilbert-pretraining-task-agnostic-visiolinguistic/). Advances in neural information processing systems, 32.
30. Zhipeng Qian, Yiwei Ma, Jiayi Ji, and Xiaoshuai Sun. 2024a. X-refseg3d: Enhancing referring 3d instance segmentation via structured cross-modal graph neural networks. In Proceedings of the AAAI conference on artificial intelligence, volume 38, pages 4551–4559. [find](https://scholar.google.com/scholar?q=Zhipeng+Qian%2C+Yiwei+Ma%2C+Jiayi+Ji%2C+and+Xiaoshuai+Sun%2E+2024a%2E+X%2Drefseg3d%3A+Enhancing+referring+3d+instance+segmentation+via+structured+cross%2Dmodal+graph+neural+networks%2E+In+Proceedings+of+the+AAAI+confer)
31. Zhipeng Qian, Yiwei Ma, Zhekai Lin, Jiayi Ji, Xiawu Zheng, Xiaoshuai Sun, and Rongrong Ji. 2024b. Multi-branch collaborative learning network for 3d visual grounding. In European Conference on Computer Vision, pages 381–398. Springer. [find](https://scholar.google.com/scholar?q=Zhipeng+Qian%2C+Yiwei+Ma%2C+Zhekai+Lin%2C+Jiayi+Ji%2C+Xiawu+Zheng%2C+Xiaoshuai+Sun%2C+and+Rongrong+Ji%2E+2024b%2E+Multi%2Dbranch+collaborative+learning+network+for+3d+visual+grounding%2E+In+European+Conference+on+Compute)
32. Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, and 1 others. 2021. [Learning transferable visual models from natural language supervision](https://md.research.elamur.ai/free/en/papers/radford-2021-learning-transferable-visual-models-natural/). In International conference on machine learning, pages 8748–8…
33. Kuniaki Saito, Kihyuk Sohn, Xiang Zhang, Chun-Liang Li, Chen-Yu Lee, Kate Saenko, and Tomas Pfister. 2023. Pic2word: Mapping pictures to words for zero-shot composed image retrieval. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19305–19314. [find](https://scholar.google.com/scholar?q=Kuniaki+Saito%2C+Kihyuk+Sohn%2C+Xiang+Zhang%2C+Chun%2DLiang+Li%2C+Chen%2DYu+Lee%2C+Kate+Saenko%2C+and+Tomas+Pfister%2E+2023%2E+Pic2word%3A+Mapping+pictures+to+words+for+zero%2Dshot+composed+image+retrieval%2E+In+Proceedings+of)
34. Shitong Sun, Fanghua Ye, and Shaogang Gong. 2023. Training-free zero-shot composed image retrieval with local concept reranking. arXiv preprint arXiv:2312.08924. — [arXiv:2312.08924](https://arxiv.org/abs/2312.08924) [find](https://scholar.google.com/scholar?q=Shitong+Sun%2C+Fanghua+Ye%2C+and+Shaogang+Gong%2E+2023%2E+Training%2Dfree+zero%2Dshot+composed+image+retrieval+with+local+concept+reranking%2E+arXiv+preprint+arXiv%3A2312%2E08924%2E+%E2%80%94+%5BarXiv%3A2312%2E08924%5D%28https%3A%2F%2Farxiv%2Eorg)
35. Zelong Sun, Dong Jing, Guoxing Yang, Nanyi Fei, and Zhiwu Lu. 2024. Leveraging large vision-language model as user intent-aware encoder for composed image retrieval. arXiv preprint arXiv:2412.11087. — [arXiv:2412.11087](https://arxiv.org/abs/2412.11087) [find](https://scholar.google.com/scholar?q=Zelong+Sun%2C+Dong+Jing%2C+Guoxing+Yang%2C+Nanyi+Fei%2C+and+Zhiwu+Lu%2E+2024%2E+Leveraging+large+vision%2Dlanguage+model+as+user+intent%2Daware+encoder+for+composed+image+retrieval%2E+arXiv+preprint+arXiv%3A2412%2E11087%2E+%E2%80%94)
36. Yucheng Suo, Fan Ma, Linchao Zhu, and Yi Yang. 2024. Knowledge-enhanced dual-stream zero-shot composed image retrieval. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 26951–26962. [find](https://scholar.google.com/scholar?q=Yucheng+Suo%2C+Fan+Ma%2C+Linchao+Zhu%2C+and+Yi+Yang%2E+2024%2E+Knowledge%2Denhanced+dual%2Dstream+zero%2Dshot+composed+image+retrieval%2E+In+Proceedings+of+the+IEEE%2FCVF+Conference+on+Computer+Vision+and+Pattern+Recogni)
37. Haomiao Tang, Jinpeng Wang, Yuang Peng, GuangHao Meng, Ruisheng Luo, Bin Chen, Long Chen, Yaowei Wang, and Shu-Tao Xia. 2025. Modeling uncertainty in composed image retrieval via probabilistic embeddings. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Vol… [find](https://scholar.google.com/scholar?q=Haomiao+Tang%2C+Jinpeng+Wang%2C+Yuang+Peng%2C+GuangHao+Meng%2C+Ruisheng+Luo%2C+Bin+Chen%2C+Long+Chen%2C+Yaowei+Wang%2C+and+Shu%2DTao+Xia%2E+2025%2E+Modeling+uncertainty+in+composed+image+retrieval+via+probabilistic+embeddi)
38. Yuanmin Tang, Xiaoting Qin, Jue Zhang, Jing Yu, Gaopeng Gou, Gang Xiong, Qingwei Ling, Saravan Rajmohan, Dongmei Zhang, and Qi Wu. 2024. Reason-before-retrieve: One-stage reflective chain-of-thoughts for training-free zero-shot composed image retrieval. arXiv preprint arXiv:2412.11077. — [arXiv:2412.11077](https://arxiv.org/abs/2412.11077) [find](https://scholar.google.com/scholar?q=Yuanmin+Tang%2C+Xiaoting+Qin%2C+Jue+Zhang%2C+Jing+Yu%2C+Gaopeng+Gou%2C+Gang+Xiong%2C+Qingwei+Ling%2C+Saravan+Rajmohan%2C+Dongmei+Zhang%2C+and+Qi+Wu%2E+2024%2E+Reason%2Dbefore%2Dretrieve%3A+One%2Dstage+reflective+chain%2Dof%2Dthoughts+)
39. Likai Tian, Jian Zhao, Zechao Hu, Zhengwei Yang, Hao Li, Lei Jin, Zheng Wang, and Xuelong Li. 2025. Ccin: Compositional conflict identification and neutralization for composed image retrieval. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 3974–3983. [find](https://scholar.google.com/scholar?q=Likai+Tian%2C+Jian+Zhao%2C+Zechao+Hu%2C+Zhengwei+Yang%2C+Hao+Li%2C+Lei+Jin%2C+Zheng+Wang%2C+and+Xuelong+Li%2E+2025%2E+Ccin%3A+Compositional+conflict+identification+and+neutralization+for+composed+image+retrieval%2E+In+Proc)
40. Lucas Ventura, Antoine Yang, Cordelia Schmid, and Gül Varol. 2024. Covr: Learning composed video retrieval from web video captions. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 5270–5279. [find](https://scholar.google.com/scholar?q=Lucas+Ventura%2C+Antoine+Yang%2C+Cordelia+Schmid%2C+and+G%C3%BCl+Varol%2E+2024%2E+Covr%3A+Learning+composed+video+retrieval+from+web+video+captions%2E+In+Proceedings+of+the+AAAI+Conference+on+Artificial+Intelligence%2C+vo)
41. Yongquan Wan, Wenhai Wang, Guobing Zou, and Bofeng Zhang. 2024. Cross-modal feature alignment and fusion for composed image retrieval. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8384–8388. [find](https://scholar.google.com/scholar?q=Yongquan+Wan%2C+Wenhai+Wang%2C+Guobing+Zou%2C+and+Bofeng+Zhang%2E+2024%2E+Cross%2Dmodal+feature+alignment+and+fusion+for+composed+image+retrieval%2E+In+Proceedings+of+the+IEEE%2FCVF+Conference+on+Computer+Vision+and+)
42. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, and 1 others. 2022. [Chain-of-thought prompting elicits reasoning in large language models](https://md.research.elamur.ai/free/en/papers/wei-2022-chain-thought-prompting-elicits-reasoning/). Advances in neural information processing systems, 35:24824– 24837.
43. Haokun Wen, Xuemeng Song, Xiaolin Chen, Yinwei Wei, Liqiang Nie, and Tat-Seng Chua. 2024. Simple but effective raw-data level multimodal fusion for composed image retrieval. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 229–… [find](https://scholar.google.com/scholar?q=Haokun+Wen%2C+Xuemeng+Song%2C+Xiaolin+Chen%2C+Yinwei+Wei%2C+Liqiang+Nie%2C+and+Tat%2DSeng+Chua%2E+2024%2E+Simple+but+effective+raw%2Ddata+level+multimodal+fusion+for+composed+image+retrieval%2E+In+Proceedings+of+the+47th)
44. Haokun Wen, Xuemeng Song, Jianhua Yin, Jianlong Wu, Weili Guan, and Liqiang Nie. 2023a. Self-training boosted multi-factor matching network for composed image retrieval. IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(5):3665– 3678. [find](https://scholar.google.com/scholar?q=Haokun+Wen%2C+Xuemeng+Song%2C+Jianhua+Yin%2C+Jianlong+Wu%2C+Weili+Guan%2C+and+Liqiang+Nie%2E+2023a%2E+Self%2Dtraining+boosted+multi%2Dfactor+matching+network+for+composed+image+retrieval%2E+IEEE+Transactions+on+Pattern+A)
45. Haokun Wen, Xian Zhang, Xuemeng Song, Yinwei Wei, and Liqiang Nie. 2023b. Target-guided composed image retrieval. In Proceedings of the 31st ACM International Conference on Multimedia, pages 915– 923. [find](https://scholar.google.com/scholar?q=Haokun+Wen%2C+Xian+Zhang%2C+Xuemeng+Song%2C+Yinwei+Wei%2C+and+Liqiang+Nie%2E+2023b%2E+Target%2Dguided+composed+image+retrieval%2E+In+Proceedings+of+the+31st+ACM+International+Conference+on+Multimedia%2C+pages+915%E2%80%93+923%2E)
46. Zhirong Wu, Yuanjun Xiong, Stella X Yu, and Dahua Lin. 2018. Unsupervised feature learning via non-parametric instance discrimination. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3733–3742. [find](https://scholar.google.com/scholar?q=Zhirong+Wu%2C+Yuanjun+Xiong%2C+Stella+X+Yu%2C+and+Dahua+Lin%2E+2018%2E+Unsupervised+feature+learning+via+non%2Dparametric+instance+discrimination%2E+In+Proceedings+of+the+IEEE+conference+on+computer+vision+and+patt)
47. Eric Xing, Pranavi Kolouju, Robert Pless, Abby Stylianou, and Nathan Jacobs. 2025. Context-cir: Learning from concepts in text for composed image retrieval. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 19638– 19648. [find](https://scholar.google.com/scholar?q=Eric+Xing%2C+Pranavi+Kolouju%2C+Robert+Pless%2C+Abby+Stylianou%2C+and+Nathan+Jacobs%2E+2025%2E+Context%2Dcir%3A+Learning+from+concepts+in+text+for+composed+image+retrieval%2E+In+Proceedings+of+the+Computer+Vision+and+P)
48. Wenjie Li, and Min Zhang. 2024b. Gme: Improving universal multimodal retrieval by multimodal llms. arXiv preprint arXiv:2412.16855. — [arXiv:2412.16855](https://arxiv.org/abs/2412.16855) [find](https://scholar.google.com/scholar?q=Wenjie+Li%2C+and+Min+Zhang%2E+2024b%2E+Gme%3A+Improving+universal+multimodal+retrieval+by+multimodal+llms%2E+arXiv+preprint+arXiv%3A2412%2E16855%2E+%E2%80%94+%5BarXiv%3A2412%2E16855%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2412%2E16855%29)
49. Zhuosheng Zhang, Aston Zhang, Mu Li, Hai Zhao, George Karypis, and Alex Smola. 2023. [Multimodal chain-of-thought reasoning in language models](https://md.research.elamur.ai/free/en/papers/zhang-2023-multimodal-chain-thought-reasoning-language/). arXiv preprint arXiv:2302.00923.
50. Yida Zhao, Yuqing Song, and Qin Jin. 2022. Progressive learning for image retrieval with hybrid-modality queries. In Proceedings of the 45th international ACM SIGIR conference on research and development in information retrieval, pages 1012–1021. [find](https://scholar.google.com/scholar?q=Yida+Zhao%2C+Yuqing+Song%2C+and+Qin+Jin%2E+2022%2E+Progressive+learning+for+image+retrieval+with+hybrid%2Dmodality+queries%2E+In+Proceedings+of+the+45th+international+ACM+SIGIR+conference+on+research+and+developm)
51. Ge Zheng, Bin Yang, Jiajin Tang, Hong-Yu Zhou, and Sibei Yang. 2023. Ddcot: Duty-distinct chain-of-thought prompting for multimodal reasoning in language models. Advances in Neural Information Processing Systems, 36:5168–5191. [find](https://scholar.google.com/scholar?q=Ge+Zheng%2C+Bin+Yang%2C+Jiajin+Tang%2C+Hong%2DYu+Zhou%2C+and+Sibei+Yang%2E+2023%2E+Ddcot%3A+Duty%2Ddistinct+chain%2Dof%2Dthought+prompting+for+multimodal+reasoning+in+language+models%2E+Advances+in+Neural+Information+Process)
52. Zexin Zheng, Huangyu Dai, Lingtao Mao, Xinyu Sun, Zihan Liang, Ben Chen, Yuqing Ding, Chenyi Lei, Wenwu Ou, Han Li, and 1 others. 2025. Onevision: An end-to-end generative framework for multi-view e-commerce vision search. arXiv preprint arXiv:2510.05759. — [arXiv:2510.05759](https://arxiv.org/abs/2510.05759) [find](https://scholar.google.com/scholar?q=Zexin+Zheng%2C+Huangyu+Dai%2C+Lingtao+Mao%2C+Xinyu+Sun%2C+Zihan+Liang%2C+Ben+Chen%2C+Yuqing+Ding%2C+Chenyi+Lei%2C+Wenwu+Ou%2C+Han+Li%2C+and+1+others%2E+2025%2E+Onevision%3A+An+end%2Dto%2Dend+generative+framework+for+multi%2Dview+e%2Dc)
53. Xinxing Xu, Yong Liu, Salman Khan, Fahad Khan, Wangmeng Zuo, Rick Siow Mong Goh, Chun-Mei Feng, and 1 others. 2024. Sentence-level prompts benefit composed image retrieval. In The Twelfth International Conference on Learning Representations. [find](https://scholar.google.com/scholar?q=Xinxing+Xu%2C+Yong+Liu%2C+Salman+Khan%2C+Fahad+Khan%2C+Wangmeng+Zuo%2C+Rick+Siow+Mong+Goh%2C+Chun%2DMei+Feng%2C+and+1+others%2E+2024%2E+Sentence%2Dlevel+prompts+benefit+composed+image+retrieval%2E+In+The+Twelfth+Internationa)
54. Danni Yang, Ruohan Dong, Jiayi Ji, Yiwei Ma, Haowei Wang, Xiaoshuai Sun, and Rongrong Ji. 2024a. Exploring phrase-level grounding with text-to-image diffusion model. In European Conference on Computer Vision, pages 161–180. Springer. [find](https://scholar.google.com/scholar?q=Danni+Yang%2C+Ruohan+Dong%2C+Jiayi+Ji%2C+Yiwei+Ma%2C+Haowei+Wang%2C+Xiaoshuai+Sun%2C+and+Rongrong+Ji%2E+2024a%2E+Exploring+phrase%2Dlevel+grounding+with+text%2Dto%2Dimage+diffusion+model%2E+In+European+Conference+on+Computer)
55. Xingyu Yang, Daqing Liu, Heng Zhang, Yong Luo, Chaoyue Wang, and Jing Zhang. 2024b. Decomposing semantic shifts for composed image retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 6576–6584. [find](https://scholar.google.com/scholar?q=Xingyu+Yang%2C+Daqing+Liu%2C+Heng+Zhang%2C+Yong+Luo%2C+Chaoyue+Wang%2C+and+Jing+Zhang%2E+2024b%2E+Decomposing+semantic+shifts+for+composed+image+retrieval%2E+In+Proceedings+of+the+AAAI+Conference+on+Artificial+Intell)
56. Jinlu Zhang, Jiji Tang, Rongsheng Zhang, Tangjie Lv, and Xiaoshuai Sun. 2025. Storyweaver: A unified world model for knowledge-enhanced story character customization. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 9951–9959. [find](https://scholar.google.com/scholar?q=Jinlu+Zhang%2C+Jiji+Tang%2C+Rongsheng+Zhang%2C+Tangjie+Lv%2C+and+Xiaoshuai+Sun%2E+2025%2E+Storyweaver%3A+A+unified+world+model+for+knowledge%2Denhanced+story+character+customization%2E+In+Proceedings+of+the+AAAI+Confer)
57. Jinlu Zhang, Yiyi Zhou, Qiancheng Zheng, Xiaoxiong Du, Gen Luo, Jun Peng, Xiaoshuai Sun, and Rongrong Ji. 2024a. Fast text-to-3d-aware face generation and manipulation via direct cross-modal mapping and geometric regularization. arXiv preprint arXiv:2403.06702. — [arXiv:2403.06702](https://arxiv.org/abs/2403.06702) [find](https://scholar.google.com/scholar?q=Jinlu+Zhang%2C+Yiyi+Zhou%2C+Qiancheng+Zheng%2C+Xiaoxiong+Du%2C+Gen+Luo%2C+Jun+Peng%2C+Xiaoshuai+Sun%2C+and+Rongrong+Ji%2E+2024a%2E+Fast+text%2Dto%2D3d%2Daware+face+generation+and+manipulation+via+direct+cross%2Dmodal+mapping+a)
58. Xin Zhang, Yanzhao Zhang, Wen Xie, Mingxin Li, Ziqi Dai, Dingkun Long, Pengjun Xie, Meishan Zhang, [find](https://scholar.google.com/scholar?q=Xin+Zhang%2C+Yanzhao+Zhang%2C+Wen+Xie%2C+Mingxin+Li%2C+Ziqi+Dai%2C+Dingkun+Long%2C+Pengjun+Xie%2C+Meishan+Zhang%2C)

*Updated 1 November 2026*