{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-10-03-007",
    "source_story_id": "tmp-story-skelewam-sparse-geometry",
    "edition_id": "mp-2026-10-03-morning-0086",
    "edition_url": "https://themachinepress.com/edition/2026-10-03",
    "position": 7,
    "story_type": "dispatch",
    "section": "robotics",
    "editorial_classification": "editorial",
    "headline": "A Sparse Skeleton Beat a Video World Model",
    "slug": "a-sparse-skeleton-beat-a-video-world-model",
    "dek": "SkeleWAM represents joints, object centers and interaction points instead of reconstructing full video.",
    "summary": "SkeleWAM represents joints, object centers and interaction points instead of reconstructing full video.",
    "body_text": "SkeleWAM builds a sparse 3D skeleton online from RGB-D observations and robot proprioception, then uses it both for action generation and future-state prediction. On LIBERO-Plus, the 57.1-million-parameter system reported 85.9 percent overall success, 3.7 percentage points above Cosmos-Policy. The finding supports compact geometric state for the tested manipulation suite; it is not a general comparison across every visual world model or physical robot.",
    "why_it_matters": "SkeleWAM represents joints, object centers and interaction points instead of reconstructing full video.",
    "limitations": [
      "The finding supports compact geometric state for the tested manipulation suite; it is not a general comparison across every visual world model or physical robot."
    ],
    "importance": 9,
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-03-007/a-sparse-skeleton-beat-a-video-world-model",
    "json_url": "https://themachinepress.com/story/mp-2026-10-03-007.json",
    "first_published_at": "2026-10-03T09:00:00.000-04:00",
    "modified_at": "2026-10-03T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-10-03-007-001",
        "text": "SkeleWAM represents joints, object centers and interaction points instead of reconstructing full video.",
        "source_ids": [
          "source-2026-10-03-007"
        ],
        "qualification": "The finding supports compact geometric state for the tested manipulation suite; it is not a general comparison across every visual world model or physical robot."
      }
    ],
    "source_ids": [
      "source-2026-10-03-007"
    ],
    "tags": [
      "robot manipulation",
      "world models",
      "3D geometry"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-10-03-007",
      "title": "arXiv preprint 2610.02120",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2610.02120",
      "canonical_url": "https://arxiv.org/abs/2610.02120",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-10-01T13:35:10.000-04:00",
      "accessed_at": "2026-10-03T08:31:37.000-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-10-03-007-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "A Sparse Skeleton Beat a Video World Model",
    "publisher": "The Machine Press",
    "published_at": "2026-10-03T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-03-007/a-sparse-skeleton-beat-a-video-world-model"
  }
}