{
  "schema_version": "paper_public_manifest_v1",
  "paper_id": "vlaw_2026_09",
  "slug": "vlaw",
  "title": "VLAW：策略与世界模型共同改进",
  "authors": [],
  "source": {
    "arxiv_id": "2602.12063",
    "pdf_url": "https://arxiv.org/pdf/2602.12063",
    "project_url": "",
    "github_url": "",
    "huggingface_url": "",
    "original_source": "https://arxiv.org/pdf/2602.12063"
  },
  "site": {
    "post_url": "/posts/vlaw",
    "canonical_url": "https://haiguangboy.com/posts/vlaw",
    "cover_image": "https://static.haiguangboy.com/papers/vlaw/cover.webp"
  },
  "taxonomy": {
    "domain": "embodied_ai",
    "track": "world_model",
    "tasks": [
      "embodied_ai",
      "world_model",
      "vla",
      "action_generation",
      "robotics",
      "state_prediction",
      "具身智能",
      "机器人",
      "世界模型",
      "vlaw",
      "机器人学习"
    ],
    "related_topics": [
      {
        "paper_id": "hydra_0_action_flow_for_generalist_world_modeling_and_control_2026_08",
        "title": "英伟达的动作流：机器人动作新标准",
        "url": "https://haiguangboy.com/posts/hydra-0-action-flow",
        "relation": "contrast",
        "summary": "仍依赖真机交互、人工复位和成功标签 contradicts 建模对象的重新定义:不是复现任务特定行为,而是建模机器人运动的后果",
        "strength": "strong"
      },
      {
        "paper_id": "pointworld_scaling_3d_world_models_for_in_the_wild_robotic_manipulation_2026_08",
        "title": "PointWorld:点流统一状态动作",
        "url": "https://haiguangboy.com/posts/pointworld_scaling_3d_world_models_for_in_the_wild_robotic_manipulation",
        "relation": "same_track",
        "summary": "PointWorld:点流统一状态动作",
        "strength": "strong"
      },
      {
        "paper_id": "rl_100_performant_robotic_manipulation_with_real_world_reinforcement_learning_2026_08",
        "title": "RL 不该从零学，该做后训练",
        "url": "https://haiguangboy.com/posts/rl-100",
        "relation": "same_track",
        "summary": "仍依赖真机交互、人工复位和成功标签 validates 自主复位仍是未解决的实践瓶颈,作者自己列为未来工作",
        "strength": "strong"
      },
      {
        "paper_id": "skild_blog_introducing_s1_in_context_learning_20260817_2026_08",
        "title": "S1:机器人的GPT-3涌现时刻",
        "url": "https://haiguangboy.com/posts/skild_blog_introducing_s1_in_context_learning_20260817",
        "relation": "same_track",
        "summary": "S1:机器人的GPT-3涌现时刻",
        "strength": "strong"
      },
      {
        "paper_id": "towards_generalizable_visually_grounded_exploration_of_household_devices_2026_09",
        "title": "VGEBench:可泛化视觉探索",
        "url": "https://haiguangboy.com/posts/towards_generalizable_visually_grounded_exploration_of_household_devices",
        "relation": "same_track",
        "summary": "VGEBench:可泛化视觉探索",
        "strength": "strong"
      },
      {
        "paper_id": "robointer15_a_holistic_intermediate_representation_suite_for_embodied_world_mode_2026_07",
        "title": "不是\"要不要世界模型\"，是\"它吐出来的东西有没有结构\"",
        "url": "https://haiguangboy.com/posts/robointer15_a_holistic_intermediate_representation_suite_for_embodied_world_mode",
        "relation": "same_track",
        "summary": "不是\"要不要世界模型\"，是\"它吐出来的东西有没有结构\"",
        "strength": "strong"
      }
    ]
  },
  "analyst_take": {
    "type": "author_opinion",
    "text": "💡 和WAM综述放在一起看\n综述提醒：世界模型追求预测准确，策略追求回报，两者目标不同。VLAW把分工落到数据上：世界模型学习成功与失败，策略学习筛选后的成功经验。共同改进，不要求两者学习相同的轨迹集合。\n\n半年后回看：在同等真机预算下，扩展任务与迭代轮次后，这种交替改进能否继续带来收益。"
  },
  "ruling": {
    "importance_score": 3.0,
    "one_sentence": "VLAW：策略与世界模型共同改进"
  },
  "asset_base_url": "https://static.haiguangboy.com/papers/vlaw",
  "assets": [
    {
      "type": "pdf_screenshot",
      "object_key": "papers/vlaw/page_01.webp",
      "content_type": "image/webp",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/vlaw/page_01.webp",
      "role": "paper_first_page",
      "size_bytes": 192590
    },
    {
      "type": "pdf_screenshot",
      "object_key": "papers/vlaw/key_figure.webp",
      "content_type": "image/webp",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/vlaw/key_figure.webp",
      "role": "method_figure",
      "size_bytes": 210724
    },
    {
      "type": "cover_image",
      "object_key": "papers/vlaw/cover.webp",
      "content_type": "image/webp",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/vlaw/cover.webp",
      "role": "post_cover",
      "size_bytes": 70362
    },
    {
      "type": "public_brief",
      "object_key": "papers/vlaw/public_brief.md",
      "content_type": "text/markdown; charset=utf-8",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/vlaw/public_brief.md",
      "role": "public_brief",
      "size_bytes": 4063
    },
    {
      "type": "public_brief",
      "object_key": "papers/vlaw/public_brief.en.md",
      "content_type": "text/markdown; charset=utf-8",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/vlaw/public_brief.en.md",
      "role": "public_brief_en",
      "size_bytes": 5176
    },
    {
      "type": "public_manifest",
      "object_key": "papers/vlaw/public_manifest.en.json",
      "content_type": "application/json; charset=utf-8",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/vlaw/public_manifest.en.json",
      "role": "public_manifest_en",
      "size_bytes": 6252
    },
    {
      "type": "public_manifest",
      "object_key": "papers/vlaw/public_manifest.json",
      "content_type": "application/json; charset=utf-8",
      "upload_status": "uploaded",
      "bucket": "paper-assets",
      "url": "https://static.haiguangboy.com/papers/vlaw/public_manifest.json",
      "role": "public_manifest",
      "size_bytes": 6841
    }
  ],
  "published_at": "2026-09-19T08:32:49+08:00",
  "created_at": "2026-09-19T08:32:49+08:00",
  "updated_at": "2026-09-19T08:33:00+08:00"
}