{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-10-14-004",
    "source_story_id": "tmp-story-loopwam-deployment-rollouts",
    "edition_id": "mp-2026-10-14-morning-0096",
    "edition_url": "https://themachinepress.com/edition/2026-10-14",
    "position": 4,
    "story_type": "dispatch",
    "section": "robotics",
    "editorial_classification": "editorial",
    "headline": "Robot Failures Stay in the Training Set",
    "slug": "robot-failures-stay-in-the-training-set",
    "dek": "LoopWAM learns from successful, weak and failed deployment rollouts instead of imitating everything or discarding mistakes.",
    "summary": "LoopWAM learns from successful, weak and failed deployment rollouts instead of imitating everything or discarding mistakes.",
    "body_text": "Real deployment logs contain successes, poor actions and failures. LoopWAM uses source-dependent noise, quality steering and weighted supervision to keep video evidence across that spectrum while limiting imitation of bad behavior. The authors report that three deployment-and-training rounds lifted Table30-V2 task success from 23.67 to 43.33 percent, while two rounds on a shirt-folding task moved success from 13.20 to 55.56 percent. Results on RoboTwin and RoboCasa are also reported. The experiments do not prove that every robot improves safely from its own failures; they show a way to learn physical consequences without treating every action as an expert example.",
    "why_it_matters": "LoopWAM learns from successful, weak and failed deployment rollouts instead of imitating everything or discarding mistakes.",
    "limitations": [
      "The experiments do not prove that every robot improves safely from its own failures; they show a way to learn physical consequences without treating every action as an expert example."
    ],
    "importance": 8,
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-14-004/robot-failures-stay-in-the-training-set",
    "json_url": "https://themachinepress.com/story/mp-2026-10-14-004.json",
    "first_published_at": "2026-10-14T09:00:00.000-04:00",
    "modified_at": "2026-10-14T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-10-14-004-001",
        "text": "LoopWAM learns from successful, weak and failed deployment rollouts instead of imitating everything or discarding mistakes.",
        "source_ids": [
          "source-2026-10-14-004"
        ],
        "qualification": "The experiments do not prove that every robot improves safely from its own failures; they show a way to learn physical consequences without treating every action as an expert example."
      }
    ],
    "source_ids": [
      "source-2026-10-14-004"
    ],
    "tags": [
      "LoopWAM",
      "deployment learning",
      "world-action models"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-10-14-004",
      "title": "arXiv preprint 2610.18873, original v1",
      "publisher": "arXiv / submitting research authors",
      "url": "https://arxiv.org/abs/2610.18873v1",
      "canonical_url": "https://arxiv.org/abs/2610.18873v1",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-10-13T13:59:56.000-04:00",
      "accessed_at": "2026-10-14T08:24:30.364-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-10-14-004-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "Robot Failures Stay in the Training Set",
    "publisher": "The Machine Press",
    "published_at": "2026-10-14T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-14-004/robot-failures-stay-in-the-training-set"
  }
}