Code is publicly available at <a href=\"https://github.com/oshapio/binding-concepts-complexity\" rel=\"nofollow\">https://github.com/oshapio/binding-concepts-complexity</a></p>\n","updatedAt":"2026-06-01T12:11:38.313Z","author":{"_id":"6520898f7bf8cc2dd28b7a9c","avatarUrl":"/avatars/87a29ba95b71ee2dce18e97aa85e17a1.svg","fullname":"Arnas Uselis","name":"Gigglingface","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":5,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.824007511138916},"editors":["Gigglingface"],"editorAvatarUrls":["/avatars/87a29ba95b71ee2dce18e97aa85e17a1.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2605.31503","authors":[{"_id":"6a1d76af808ddbc3c7d43860","user":{"_id":"6520898f7bf8cc2dd28b7a9c","avatarUrl":"/avatars/87a29ba95b71ee2dce18e97aa85e17a1.svg","isPro":false,"fullname":"Arnas Uselis","user":"Gigglingface","type":"user","name":"Gigglingface"},"name":"Arnas Uselis","status":"claimed_verified","statusLastChangedAt":"2026-06-01T12:50:17.288Z","hidden":false},{"_id":"6a1d76af808ddbc3c7d43861","name":"Darina Koishigarina","hidden":false},{"_id":"6a1d76af808ddbc3c7d43862","name":"Seong Joon Oh","hidden":false}],"publishedAt":"2026-05-29T00:00:00.000Z","submittedOnDailyAt":"2026-06-01T00:00:00.000Z","title":"How can embedding models bind concepts?","submittedOnDailyBy":{"_id":"6520898f7bf8cc2dd28b7a9c","avatarUrl":"/avatars/87a29ba95b71ee2dce18e97aa85e17a1.svg","isPro":false,"fullname":"Arnas Uselis","user":"Gigglingface","type":"user","name":"Gigglingface"},"summary":"Humans easily determine which color belongs to which shape in multi-object scenes, an ability known as concept binding. Vision-language embedding models such as CLIP struggle with binding: they recognize individual concepts but fail to represent which concepts form which objects. Although CLIP behaves like a bag-of-concepts model in cross-modal retrieval, object information is recoverable from its image and text embeddings separately. We study this tension through the binding function, which maps concepts to scene embeddings. We find that scene embeddings decompose additively into object representations, explaining why uni-modal probes can recover object information. However, CLIP's binding function is high-complexity, which likely prevents the image and text encoders from learning a shared binding mechanism that generalizes to unseen concept combinations. We then ask whether this limitation is fundamental. We show that it is not. In controlled transformer models trained from scratch, binding generalization emerges with sufficient data coverage. These models learn low-complexity binding functions characterized by multiplicative interactions between concepts, enabling systematic generalization. Code is publicly available at https://github.com/oshapio/binding-concepts-complexity.","upvotes":4,"discussionId":"6a1d76af808ddbc3c7d43863","githubRepo":"https://github.com/oshapio/binding-concepts-complexity","githubRepoAddedBy":"user","ai_summary":"Vision-language models like CLIP struggle with concept binding despite recognizing individual concepts, but controlled transformer models can learn low-complexity binding functions that generalize better through multiplicative interactions.","ai_keywords":["vision-language embedding models","CLIP","concept binding","cross-modal retrieval","scene embeddings","object representations","binding function","transformer models","multiplicative interactions","generalization"],"githubStars":0},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"6520898f7bf8cc2dd28b7a9c","avatarUrl":"/avatars/87a29ba95b71ee2dce18e97aa85e17a1.svg","isPro":false,"fullname":"Arnas Uselis","user":"Gigglingface","type":"user"},{"_id":"66deb633c35391da4334f6fb","avatarUrl":"/avatars/3b2f13d4c52ead84af46ad58d6e36f16.svg","isPro":false,"fullname":"Darina Koishigarina","user":"dariina","type":"user"},{"_id":"661ab1f1fa3b144a381fa454","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/661ab1f1fa3b144a381fa454/IlpZBb9NCjo7ntFwMIH53.png","isPro":false,"fullname":"Urro","user":"urroxyz","type":"user"},{"_id":"638a50450f10aa3064f03f23","avatarUrl":"/avatars/0c068458e42950c851758a238225c3a6.svg","isPro":false,"fullname":"Seong Joon Oh","user":"coallaoh","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2605/2605.31503.md"}">
How can embedding models bind concepts?
Abstract
Vision-language models like CLIP struggle with concept binding despite recognizing individual concepts, but controlled transformer models can learn low-complexity binding functions that generalize better through multiplicative interactions.
AI-generated summary
Humans easily determine which color belongs to which shape in multi-object scenes, an ability known as concept binding. Vision-language embedding models such as CLIP struggle with binding: they recognize individual concepts but fail to represent which concepts form which objects. Although CLIP behaves like a bag-of-concepts model in cross-modal retrieval, object information is recoverable from its image and text embeddings separately. We study this tension through the binding function, which maps concepts to scene embeddings. We find that scene embeddings decompose additively into object representations, explaining why uni-modal probes can recover object information. However, CLIP's binding function is high-complexity, which likely prevents the image and text encoders from learning a shared binding mechanism that generalizes to unseen concept combinations. We then ask whether this limitation is fundamental. We show that it is not. In controlled transformer models trained from scratch, binding generalization emerges with sufficient data coverage. These models learn low-complexity binding functions characterized by multiplicative interactions between concepts, enabling systematic generalization. Code is publicly available at https://github.com/oshapio/binding-concepts-complexity.
Community
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2605.31503 in a model README.md to link it from this page.
Cite arxiv.org/abs/2605.31503 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2605.31503 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.