[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-94361":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":9,"languages":9,"totalLinesOfCode":9,"stars":10,"forks":11,"watchers":12,"openIssues":13,"contributorsCount":11,"subscribersCount":11,"size":11,"stars1d":11,"stars7d":11,"stars30d":14,"stars90d":11,"forks30d":11,"starsTrendScore":11,"compositeScore":15,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":16,"fork":16,"defaultBranch":17,"hasWiki":18,"hasPages":16,"topics":19,"createdAt":9,"pushedAt":9,"updatedAt":20,"readmeContent":21,"aiSummary":22,"trendingCount":11,"starSnapshotCount":11,"syncStatus":14,"lastSyncTime":23,"discoverSource":24},94361,"ShadowDancer","AlayaLab\u002FShadowDancer","AlayaLab","Code release for https:\u002F\u002Fshadowdancer-1.github.io\u002F",null,114,0,7,1,2,37.2,false,"main",true,[],"2026-08-24 04:01:22","\u003Ch1 align=\"center\">\n  \u003Cimg src=\"assets\u002Ficon2.png\" height=\"60\" align=\"center\" alt=\"\" \u002F> ShadowDancer\n\u003C\u002Fh1>\n\n\u003Cp align=\"center\">\u003Cb>Teaching Video World Models Any Action from a Video and Its Shadow\u003C\u002Fb>\u003C\u002Fp>\n\n\u003Cdiv align=\"center\">\n\n[![Website](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FWebsite-ShadowDancer-blue)](https:\u002F\u002Fshadowdancer-1.github.io\u002F)\n[![arXiv](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FarXiv-2607.28362-red)](https:\u002F\u002Farxiv.org\u002Fabs\u002F2607.28362)\n\u003C!-- [![HuggingFace](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002F🤗%20Weights-ShadowDancer-yellow)](TODO_hf_url) -->\n\n\u003C\u002Fdiv>\n\n> #### [ShadowDancer: Teaching Video World Models Any Action by Learning Unified Dynamics Representations from a Video and Its Shadow](https:\u002F\u002Farxiv.org\u002Fabs\u002F2607.28362)\n>\n> ##### [Jin Cao](https:\u002F\u002Fjin-cao-tma.github.io\u002F), Zian Meng, [Kaipeng Zhang](https:\u002F\u002Fkpzhang93.github.io\u002F)†  († corresponding author)\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"assets\u002Fteaser.png\" width=\"100%\" alt=\"ShadowDancer teaser\" \u002F>\n\u003C\u002Fp>\n\nShadowDancer gives interactive video world models an **any-action, frame-level control**.\n\nThe code and model weights are being cleaned up and are undergoing internal review before release.🚧 \n\n## Citation\n\n```bibtex\n@misc{cao2026shadow,\n  title={ShadowDancer: Teaching Video World Models Any Action by Learning Unified Dynamics Representations from a Video and Its Shadow},\n  author={Jin Cao and Zian Meng and Kaipeng Zhang},\n  year={2026},\n  eprint={2607.28362},\n  archivePrefix={arXiv},\n  primaryClass={cs.CV},\n  url={https:\u002F\u002Farxiv.org\u002Fabs\u002F2607.28362},\n}\n```\n","ShadowDancer 是一个面向视频世界模型（Video World Models）的动作控制方法，旨在通过单个原始视频及其对应的“影子”（shadow）视频，学习统一的动力学表征，从而实现对模型的任意动作指令与帧级精确控制。其核心技术基于对比学习与隐空间动力学解耦，无需额外动作标注或物理仿真器，支持零样本动作泛化。适用于需要细粒度、可编辑视频生成与交互式模拟的场景，如机器人动作规划预演、教育类动画生成、可控视频编辑等研究方向。","2026-08-07 02:30:08","CREATED_QUERY"]