{"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/mind-the-gap-understanding-the-modality-gap","title":"Mind the Gap: Understanding the Modality Gap in Multi-modal Contrastive Representation Learning","arxiv_id":"2203.02053","date":"2022-03-03","proceeding":null,"authors":["Weixin Liang","Yuhui Zhang","Yongchan Kwon","Serena Yeung","James Zou"],"abstract":"We present modality gap, an intriguing geometric phenomenon of the representation space of multi-modal models. Specifically, we show that different data modalities (e.g. images and text) are embedded at arm's length in their shared representation in multi-modal models such as CLIP. Our systematic analysis demonstrates that this gap is caused by a combination of model initialization and contrastive learning optimization. In model initialization, we show empirically and theoretically that the representation of a common deep neural network is restricted to a narrow cone. As a consequence, in a multi-modal model with two encoders, the representations of the two modalities are clearly apart when the model is initialized. During optimization, contrastive learning keeps the different modalities separate by a certain distance, which is influenced by the temperature parameter in the loss function. Our experiments further demonstrate that varying the modality gap distance has a significant impact in improving the model's downstream zero-shot classification performance and fairness. Our code and data are available at https://modalitygap.readthedocs.io/","url_abs":"https://arxiv.org/abs/2203.02053v2","url_pdf":"https://arxiv.org/pdf/2203.02053v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"mind-the-gap-understanding-the-modality-gap","repo_url":"https://github.com/weixin-liang/modality-gap","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"mind-the-gap-understanding-the-modality-gap","repo_url":"https://github.com/megvii-research/protoclip","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"contrastive-learning","task_name":"Contrastive Learning"},{"task_slug":"fairness","task_name":"Fairness"},{"task_slug":"representation-learning","task_name":"Representation Learning"},{"task_slug":"zero-shot-learning","task_name":"Zero-Shot Learning"},{"task_slug":null,"task_name":"zero-shot-classification"}],"methods":[{"method_slug":"clip","method_name":"CLIP"},{"method_slug":"contrastive-learning","method_name":"Contrastive Learning"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2203.02053","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2203.02053"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+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":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/weixin-liang/modality-gap","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/megvii-research/protoclip","reach":{"status":"ok","spdx":"NOASSERTION"}}],"summary":{"unverified":5},"by_repo_kind":{"official":{"samples":5,"ran":0,"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":0,"samples":[{"code_sha256_prefix":"593596c1e55f3c89","entry":"align_loss_","repo":"weixin-liang/modality-gap","repo_kind":"official","path":"Table_1_Implications_CLIP_Zero_Shot/training/utils.py","file_url":"https://github.com/weixin-liang/modality-gap/blob/HEAD/Table_1_Implications_CLIP_Zero_Shot/training/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"593596c1e55f3c89"}},{"code_sha256_prefix":"35e7c386f1376a2a","entry":"encode_clip","repo":"weixin-liang/modality-gap","repo_kind":"official","path":"Table_1_Implications_CLIP_Zero_Shot/training/utils.py","file_url":"https://github.com/weixin-liang/modality-gap/blob/HEAD/Table_1_Implications_CLIP_Zero_Shot/training/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"35e7c386f1376a2a"}},{"code_sha256_prefix":"604cad9154e2606d","entry":"evaluate_binary_classification","repo":"weixin-liang/modality-gap","repo_kind":"official","path":"Table_1_Implications_CLIP_Zero_Shot/training/train_clip.py","file_url":"https://github.com/weixin-liang/modality-gap/blob/HEAD/Table_1_Implications_CLIP_Zero_Shot/training/train_clip.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"604cad9154e2606d"}},{"code_sha256_prefix":"639e446fd84eec00","entry":"evaluate_classification","repo":"weixin-liang/modality-gap","repo_kind":"official","path":"Table_1_Implications_CLIP_Zero_Shot/training/train_clip.py","file_url":"https://github.com/weixin-liang/modality-gap/blob/HEAD/Table_1_Implications_CLIP_Zero_Shot/training/train_clip.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"639e446fd84eec00"}},{"code_sha256_prefix":"cb1878b5f902b937","entry":"evaluate_retrieval","repo":"weixin-liang/modality-gap","repo_kind":"official","path":"Table_1_Implications_CLIP_Zero_Shot/training/train_clip.py","file_url":"https://github.com/weixin-liang/modality-gap/blob/HEAD/Table_1_Implications_CLIP_Zero_Shot/training/train_clip.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"cb1878b5f902b937"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}