[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag-world-models":3},{"tag":4,"articles":11},{"id":5,"name":6,"slug":7,"article_count":8,"description_zh":9,"description_en":10},"7abda745-cf3a-41af-9493-634a33a600ce","world models","world-models",4,"世界模型指的是能從觀測中學習環境動態、再用來做預測與規劃的模型，常見於強化學習、機器人控制與影片理解。它之所以重要，在於能把 AI 從單次生成推向可推演、可決策的系統，並降低長期規劃與多代理互動的成本。","World models learn environment dynamics from observations so agents can predict outcomes and plan ahead. They matter in reinforcement learning, robotics, and video understanding, where latent planning, long-horizon control, and multi-agent action binding can reduce compute and improve decision quality.",[12,21,28],{"id":13,"slug":14,"title":15,"summary":16,"category":17,"image_url":18,"cover_image":18,"language":19,"created_at":20},"5e1d7109-dea2-4002-8e74-bf6331f46c05","hierarchical-planning-latent-world-models-en","Hierarchical Planning Cuts World-Model Search Cost","A hierarchical latent world-model planner improves long-horizon control and cuts planning compute, with zero-shot gains on real robots.","research","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775455787999-4nzg.png","en","2026-04-06T06:09:31.947292+00:00",{"id":22,"slug":23,"title":24,"summary":25,"category":17,"image_url":26,"cover_image":26,"language":19,"created_at":27},"cb4a3bf4-e0a1-40e3-a421-112aa0d632e9","actionparty-multi-subject-action-binding-en","ActionParty binds actions to multiple agents","ActionParty tackles multi-agent control in video world models by binding actions to subjects, with reported gains in action-following and identity consistency.","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775196227996-iurw.png","2026-04-03T06:03:31.477984+00:00",{"id":29,"slug":30,"title":31,"summary":32,"category":33,"image_url":34,"cover_image":34,"language":19,"created_at":35},"15c2f00f-4c48-4580-a13e-74626eb520f7","five-ai-infra-frontiers-bessemer-2026-en","Five AI Infra Frontiers Bessemer Expects for 2026","Bessemer’s 2026 AI infra roadmap points to memory, continual learning, RL, inference, and world models as the next big build areas.","industry","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775164380914-xfye.png","2026-04-02T21:12:40.223864+00:00"]