{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-10-14-016",
    "source_story_id": "tmp-brief-data-free-lattice-diffusion",
    "edition_id": "mp-2026-10-14-morning-0096",
    "edition_url": "https://themachinepress.com/edition/2026-10-14",
    "position": 18,
    "story_type": "ticker",
    "section": "research",
    "editorial_classification": "editorial",
    "headline": "A Diffusion Sampler Trains on Its Own Field Configurations",
    "slug": "a-diffusion-sampler-trains-on-its-own-field-configurations",
    "dek": "A target-score identity removes the need for reference samples in a lattice-field experiment.",
    "summary": "A target-score identity removes the need for reference samples in a lattice-field experiment.",
    "body_text": "The authors train diffusion models from their own generated samples when the unnormalized target density is available. In two-dimensional phi-four theory, selected importance-weighted and fixed-point variants matched training based on Hamiltonian Monte Carlo reference data. The claim is limited to the evaluated field theory.",
    "why_it_matters": "A target-score identity removes the need for reference samples in a lattice-field experiment.",
    "limitations": [],
    "importance": 6,
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-14-016/a-diffusion-sampler-trains-on-its-own-field-configurations",
    "json_url": "https://themachinepress.com/story/mp-2026-10-14-016.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-2026-10-14-016-001",
        "text": "A target-score identity removes the need for reference samples in a lattice-field experiment.",
        "source_ids": [
          "source-2026-10-14-018"
        ],
        "qualification": null
      }
    ],
    "source_ids": [
      "source-2026-10-14-018"
    ],
    "tags": [
      "lattice field theory",
      "diffusion models",
      "sampling"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-10-14-018",
      "title": "arXiv preprint 2610.18863, original v1",
      "publisher": "arXiv / submitting research authors",
      "url": "https://arxiv.org/abs/2610.18863v1",
      "canonical_url": "https://arxiv.org/abs/2610.18863v1",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-10-13T13:59:33.000-04:00",
      "accessed_at": "2026-10-14T08:24:30.364-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-10-14-016-001"
      ]
    }
  ],
  "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 Diffusion Sampler Trains on Its Own Field Configurations",
    "publisher": "The Machine Press",
    "published_at": "2026-10-14T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-10-14-016/a-diffusion-sampler-trains-on-its-own-field-configurations"
  }
}