{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-08-18-011",
    "source_story_id": "tmp-story-flow-face-model-inversion",
    "edition_id": "mp-2026-08-18-morning-0040",
    "edition_url": "https://themachinepress.com/edition/2026-08-18",
    "position": 11,
    "story_type": "dispatch",
    "section": "safety-security",
    "editorial_classification": "editorial",
    "headline": "The Attack Steered Noise Toward a Face Model’s Memory",
    "slug": "the-attack-steered-noise-toward-a-face-model-s-memory",
    "dek": "A white-box inversion method injects identity gradients through a flow-matching trajectory to reconstruct representative target-class faces.",
    "summary": "A white-box inversion method injects identity gradients through a flow-matching trajectory to reconstruct representative target-class faces.",
    "body_text": "SFMI first trains an unconditional flow-matching prior over faces, then backpropagates through the target recognition model to guide intermediate samples toward a selected identity class. Under an identity-disjoint CelebA evaluation against ArcFace, the paper reports 0.9248 attack accuracy, FID 22.61 and LPIPS 0.3874, with competitive results across additional targets. The images are representative reconstructions rather than recovered source photographs, but the experiment illustrates the privacy exposure created by white-box access to recognition models.",
    "why_it_matters": "A white-box inversion method injects identity gradients through a flow-matching trajectory to reconstruct representative target-class faces.",
    "limitations": [],
    "importance": 9,
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-18-011/the-attack-steered-noise-toward-a-face-model-s-memory",
    "json_url": "https://themachinepress.com/story/mp-2026-08-18-011.json",
    "first_published_at": "2026-08-18T09:00:00.000-04:00",
    "modified_at": "2026-08-18T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-08-18-011-001",
        "text": "A white-box inversion method injects identity gradients through a flow-matching trajectory to reconstruct representative target-class faces.",
        "source_ids": [
          "source-2026-08-18-011"
        ],
        "qualification": null
      }
    ],
    "source_ids": [
      "source-2026-08-18-011"
    ],
    "tags": [
      "model inversion",
      "face recognition",
      "privacy",
      "flow matching"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-08-18-011",
      "title": "arXiv preprint 2608.16791",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2608.16791",
      "canonical_url": "https://arxiv.org/abs/2608.16791",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-08-16T20:00:00.000-04:00",
      "accessed_at": "2026-08-18T08:25:00.000-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-08-18-011-001"
      ]
    }
  ],
  "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 Attack Steered Noise Toward a Face Model’s Memory",
    "publisher": "The Machine Press",
    "published_at": "2026-08-18T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-18-011/the-attack-steered-noise-toward-a-face-model-s-memory"
  }
}