• We propose MetaphorVU-Bench, which is the first benchmark dedicated to systematic and comprehensive evaluation for metaphorical video understanding.<br>• We conduct extensive experiments and analysis, reveal\u0002ing the deficiencies of current MLLMs and providing insights into the underlying causes of their failures.<br>• We construct MetaphorBoost, boosting metaphorical video understanding via inference-time mapping aug\u0002mentation based on a metaphorical knowledge graph.</p>\n<p>ICML 2026 spotlight</p>\n","updatedAt":"2026-05-26T01:47:19.591Z","author":{"_id":"63664c8fa2abcdf2fd6425ed","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63664c8fa2abcdf2fd6425ed/IywpB0DXZ_twkmZmVSCCD.jpeg","fullname":"Li Zhuoqun","name":"lzq2021","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8497735857963562},"editors":["lzq2021"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/63664c8fa2abcdf2fd6425ed/IywpB0DXZ_twkmZmVSCCD.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2605.25461","authors":[{"_id":"6a14fa8eb57a1823d57089ec","name":"Zhuoqun Li","hidden":false},{"_id":"6a14fa8eb57a1823d57089ed","name":"Boxi Cao","hidden":false},{"_id":"6a14fa8eb57a1823d57089ee","name":"Guiping Jiang","hidden":false},{"_id":"6a14fa8eb57a1823d57089ef","name":"Fangrui Lv","hidden":false},{"_id":"6a14fa8eb57a1823d57089f0","name":"Ruotong Pan","hidden":false},{"_id":"6a14fa8eb57a1823d57089f1","name":"Jianan Wang","hidden":false},{"_id":"6a14fa8eb57a1823d57089f2","name":"Xiangyu Wu","hidden":false},{"_id":"6a14fa8eb57a1823d57089f3","name":"Hongyu Lin","hidden":false},{"_id":"6a14fa8eb57a1823d57089f4","name":"Yaojie Lu","hidden":false},{"_id":"6a14fa8eb57a1823d57089f5","name":"Yong Du","hidden":false},{"_id":"6a14fa8eb57a1823d57089f6","name":"Ruyin Jia","hidden":false},{"_id":"6a14fa8eb57a1823d57089f7","name":"Liyan","hidden":false},{"_id":"6a14fa8eb57a1823d57089f8","name":"Tingting Gao","hidden":false},{"_id":"6a14fa8eb57a1823d57089f9","name":"Han Li","hidden":false},{"_id":"6a14fa8eb57a1823d57089fa","name":"Xianpei Han","hidden":false},{"_id":"6a14fa8eb57a1823d57089fb","name":"Le Sun","hidden":false}],"mediaUrls":["https://cdn-uploads.huggingface.co/production/uploads/63664c8fa2abcdf2fd6425ed/Tswo8M6WoMJBnuTAmj03V.png"],"publishedAt":"2026-05-25T00:00:00.000Z","submittedOnDailyAt":"2026-05-26T00:00:00.000Z","title":"MetaphorVU: Towards Metaphorical Video Understanding","submittedOnDailyBy":{"_id":"63664c8fa2abcdf2fd6425ed","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63664c8fa2abcdf2fd6425ed/IywpB0DXZ_twkmZmVSCCD.jpeg","isPro":false,"fullname":"Li Zhuoqun","user":"lzq2021","type":"user","name":"lzq2021"},"summary":"Metaphorical videos are prevalent across various real-world scenarios to convey complex ideas, and understanding them typically requires high-order cognitive capabilities. The lack of systematic studies on metaphorical video understanding not only constrains the real-world applicability of MLLMs but also impedes the thorough assessment of their high-order cognitive capabilities. To bridge this gap, we propose MetaphorVU-Bench, the first systematic and comprehensive benchmark dedicated to metaphorical video understanding. Through experiments, we find current MLLMs struggle with accurate metaphorical video understanding, lagging far behind human level, primarily due to defective cross-domain mapping. Motivated by this finding, we construct a metaphor knowledge graph as mapping augmentation and propose MetaphorBoost, an inference-time enhancement framework achieving consistent performance improvement. Our benchmark, analysis, and method provide useful insights and a foundation for future research on advancing MLLMs.","upvotes":4,"discussionId":"6a14fa8eb57a1823d57089fc","projectPage":"https://huggingface.co/datasets/lzq2021/MetaphorVU-Bench","githubRepo":"https://github.com/icip-cas/MetaphorVU","githubRepoAddedBy":"user","ai_summary":"Current multimodal large language models struggle with metaphorical video understanding due to poor cross-domain mapping, prompting the development of a new benchmark and enhancement framework.","ai_keywords":["multimodal large language models","metaphorical video understanding","cross-domain mapping","metaphor knowledge graph","MetaphorBoost","inference-time enhancement"],"githubStars":3},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"63664c8fa2abcdf2fd6425ed","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63664c8fa2abcdf2fd6425ed/IywpB0DXZ_twkmZmVSCCD.jpeg","isPro":false,"fullname":"Li Zhuoqun","user":"lzq2021","type":"user"},{"_id":"68d66ad3e690f3f546768c50","avatarUrl":"/avatars/cdd80f9a76c0e433d6685725b80aafb3.svg","isPro":false,"fullname":"Ke Bao","user":"ispobock1","type":"user"},{"_id":"5fc052241160c47d1d438556","avatarUrl":"/avatars/f508707fc92f1b42f7897b12b727754c.svg","isPro":false,"fullname":"Boxi Cao","user":"Bowieee","type":"user"},{"_id":"643b62ac065961b2252abb7a","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/643b62ac065961b2252abb7a/6F140r2GfdpItuvSk8GNc.jpeg","isPro":false,"fullname":"zuijiang","user":"zuijiang","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2605/2605.25461.md"}">

Get this paper in your agent:

Don't have the latest CLI?

Cite arxiv.org/abs/2605.25461 in a model README.md to link it from this page.

Cite arxiv.org/abs/2605.25461 in a Space README.md to link it from this page.

Add this paper to a collection to link it from this page.