{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-10-30-010",
    "source_story_id": "tmp-story-physvlmm-physical-belief-maps",
    "edition_id": "mp-2026-10-30-morning-0112",
    "edition_url": "https://themachinepress.com/edition/2026-10-30",
    "position": 10,
    "story_type": "dispatch",
    "section": "robotics",
    "editorial_classification": "editorial",
    "headline": "The Robot Map Learns How a Drawer Will Pull",
    "slug": "the-robot-map-learns-how-a-drawer-will-pull",
    "dek": "Across doors and drawers in 18 homes, physical belief maps identified joint type in 86 percent of held-out pulls and answered in 13 milliseconds per case.",
    "summary": "Across doors and drawers in 18 homes, physical belief maps identified joint type in 86 percent of held-out pulls and answered in 13 milliseconds per case.",
    "body_text": "A semantic map can label a door without knowing how that door will move under force. PhysVLMM stores a probabilistic physical belief for each articulated part, updates it after measured interactions and shares evidence with visually similar untouched parts. Across doors and drawers in 18 homes, it identified joint type in 86 percent of held-out pulls, versus 48 percent for a learned single-view method. Response prediction had 41 percent lower error than differentiable simulation and took 13 milliseconds instead of 230 seconds per case. The method can answer at a stated risk or abstain, with reported false-safe rates staying within requested risk levels. The study covers the tested household fixtures and actions, not every articulated object.",
    "why_it_matters": "Across doors and drawers in 18 homes, physical belief maps identified joint type in 86 percent of held-out pulls and answered in 13 milliseconds per case.",
    "limitations": [
      "The study covers the tested household fixtures and actions, not every articulated object."
    ],
    "importance": 9,
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-30-010/the-robot-map-learns-how-a-drawer-will-pull",
    "json_url": "https://themachinepress.com/story/mp-2026-10-30-010.json",
    "first_published_at": "2026-10-30T09:00:00.000-04:00",
    "modified_at": "2026-10-30T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-10-30-010-001",
        "text": "Across doors and drawers in 18 homes, physical belief maps identified joint type in 86 percent of held-out pulls and answered in 13 milliseconds per case.",
        "source_ids": [
          "source-2026-10-30-010"
        ],
        "qualification": "The study covers the tested household fixtures and actions, not every articulated object."
      }
    ],
    "source_ids": [
      "source-2026-10-30-010"
    ],
    "tags": [
      "robotics",
      "scene maps",
      "uncertainty",
      "articulated objects"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-10-30-010",
      "title": "arXiv preprint 2610.40539, original v1",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2610.40539v1",
      "canonical_url": "https://arxiv.org/abs/2610.40539v1",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-10-28T20:00:00.000-04:00",
      "accessed_at": "2026-10-30T08:14:40.422-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-10-30-010-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "The Robot Map Learns How a Drawer Will Pull",
    "publisher": "The Machine Press",
    "published_at": "2026-10-30T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-30-010/the-robot-map-learns-how-a-drawer-will-pull"
  }
}