{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/arxiv-2601-03728","title":"CSMCIR: CoT-Enhanced Symmetric Alignment with Memory Bank for Composed Image Retrieval","arxiv_id":"2601.03728","date":"2026-01-07","proceeding":null,"authors":["Zhipeng Qian","Zihan Liang","Yufei Ma","Ben Chen","Huangyu Dai","Yiwei Ma","Jiayi Ji","Chenyi Lei","Han Li","Xiaoshuai Sun"],"abstract":"Composed Image Retrieval (CIR) enables users to search for target images using both a reference image and manipulation text, offering substantial advantages over single-modality retrieval systems. However, existing CIR methods suffer from representation space fragmentation: queries and targets comprise heterogeneous modalities and are processed by distinct encoders, forcing models to bridge misaligned representation spaces only through post-hoc alignment, which fundamentally limits retrieval performance. This architectural asymmetry manifests as three distinct, well-separated clusters in the feature space, directly demonstrating how heterogeneous modalities create fundamentally misaligned representation spaces from initialization. In this work, we propose CSMCIR, a unified representation framework that achieves efficient query-target alignment through three synergistic components. First, we introduce a Multi-level Chain-of-Thought (MCoT) prompting strategy that guides Multimodal Large Language Models to generate discriminative, semantically compatible captions for target images, establishing modal symmetry. Building upon this, we design a symmetric dual-tower architecture where both query and target sides utilize the identical shared-parameter Q-Former for cross-modal encoding, ensuring consistent feature representations and further reducing the alignment gap. Finally, this architectural symmetry enables an entropy-based, temporally dynamic Memory Bank strategy that provides high-quality negative samples while maintaining consistency with the evolving model state. Extensive experiments on four benchmark datasets demonstrate that our CSMCIR achieves state-of-the-art performance with superior training efficiency. Comprehensive ablation studies further validate the effectiveness of each proposed component.","url_abs":"https://arxiv.org/abs/2601.03728","url_pdf":"https://arxiv.org/pdf/2601.03728","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2601.03728","atlas_url":null,"mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2601.03728"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/qzp2018/CSMCIR","reach":{"status":"ok"}}],"summary":{"ran_honours":1,"ran_fixture":1,"ran_draft_wrong":2,"ran_violates":1,"unverified":4},"by_repo_kind":{"found_in_text":{"samples":9,"ran":5,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":9,"samples":[{"code_sha256_prefix":"b6c1f837e6faaeba","entry":"collate_fn","repo":"qzp2018/CSMCIR","repo_kind":"found_in_text","path":"src/data_utils_csmcir.py","file_url":"https://github.com/qzp2018/CSMCIR/blob/HEAD/src/data_utils_csmcir.py","link_basis":"plan_row","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"b6c1f837e6faaeba"}},{"code_sha256_prefix":"7d214655954e2bc6","entry":"getAttMap","repo":"qzp2018/CSMCIR","repo_kind":"found_in_text","path":"src/lavis/common/gradcam.py","file_url":"https://github.com/qzp2018/CSMCIR/blob/HEAD/src/lavis/common/gradcam.py","link_basis":"plan_row","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"7d214655954e2bc6"}},{"code_sha256_prefix":"98589643273920ba","entry":"main_process","repo":"qzp2018/CSMCIR","repo_kind":"found_in_text","path":"src/lavis/common/dist_utils.py","file_url":"https://github.com/qzp2018/CSMCIR/blob/HEAD/src/lavis/common/dist_utils.py","link_basis":"plan_row","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"98589643273920ba"}},{"code_sha256_prefix":"791c72070c0b1cfb","entry":"node_to_dict","repo":"qzp2018/CSMCIR","repo_kind":"found_in_text","path":"src/lavis/common/config.py","file_url":"https://github.com/qzp2018/CSMCIR/blob/HEAD/src/lavis/common/config.py","link_basis":"plan_row","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"791c72070c0b1cfb"}},{"code_sha256_prefix":"7c508037b40522af","entry":"str2bool","repo":"qzp2018/CSMCIR","repo_kind":"found_in_text","path":"src/cirr_test_submission_csmcir.py","file_url":"https://github.com/qzp2018/CSMCIR/blob/HEAD/src/cirr_test_submission_csmcir.py","link_basis":"harvester_set","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"7c508037b40522af"}},{"code_sha256_prefix":"56e33e39141001a3","entry":"calculate_blip_parameters","repo":"qzp2018/CSMCIR","repo_kind":"found_in_text","path":"src/blip_fine_tune_2_csmcir.py","file_url":"https://github.com/qzp2018/CSMCIR/blob/HEAD/src/blip_fine_tune_2_csmcir.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"56e33e39141001a3"}},{"code_sha256_prefix":"643539116e049080","entry":"download_cached_file","repo":"qzp2018/CSMCIR","repo_kind":"found_in_text","path":"src/lavis/common/dist_utils.py","file_url":"https://github.com/qzp2018/CSMCIR/blob/HEAD/src/lavis/common/dist_utils.py","link_basis":"plan_row","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"643539116e049080"}},{"code_sha256_prefix":"65021e0b25e8fe04","entry":"squarepad_transform","repo":"qzp2018/CSMCIR","repo_kind":"found_in_text","path":"src/data_utils_csmcir.py","file_url":"https://github.com/qzp2018/CSMCIR/blob/HEAD/src/data_utils_csmcir.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"65021e0b25e8fe04"}},{"code_sha256_prefix":"bcd92b2283e5bb7e","entry":"targetpad_transform","repo":"qzp2018/CSMCIR","repo_kind":"found_in_text","path":"src/data_utils_csmcir.py","file_url":"https://github.com/qzp2018/CSMCIR/blob/HEAD/src/data_utils_csmcir.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"bcd92b2283e5bb7e"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.CV","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}