{
  "schema_version": "1.1",
  "id": "s13:https://arxiv.org/abs/2611.06864v1",
  "slug": "2611-06864v1-1huxo8d",
  "url": "https://feed7.dev/p/2611-06864v1-1huxo8d",
  "title": "CounterFactual Retrieval for Evidence-Aware Individual Animal Identification",
  "why_included": "CFR tests whether animal identification models rely on the animal or on background cues by comparing retrieval across full-image, foreground-only, and shortcut-only views.",
  "summary": "The **Counterfactual Foreground Retrieval (CFR)** protocol evaluates the same identification task on full images, animal-only foregrounds, and shortcut-only views. The authors release around **0.5M masks across 34 datasets**, plus evaluation and mask-generation code.",
  "practical_implication": "For image retrieval systems, add evidence checks alongside accuracy: compare performance with the subject isolated and removed to see whether background cues carry identity signal.",
  "agent_context": "The **Counterfactual Foreground Retrieval (CFR)** protocol evaluates the same identification task on full images, animal-only foregrounds, and shortcut-only views. The authors release around **0.5M masks across 34 datasets**, plus evaluation and mask-generation code.\n\nFor image retrieval systems, add evidence checks alongside accuracy: compare performance with the subject isolated and removed to see whether background cues carry identity signal.\n\nThe work concerns individual animal identification. The material does not report specific results for other retrieval domains or establish that the protocol transfers unchanged.",
  "source": {
    "name": "arXiv",
    "url": "https://arxiv.org/abs/2611.06864v1",
    "published_at": "2026-11-04T18:19:26.000Z"
  },
  "source_class": "blog_post",
  "content_type": "Paper",
  "layer": "benchmark",
  "domains": [
    "image"
  ],
  "topics": [
    "agent-evals",
    "benchmark-integrity"
  ],
  "verification": {
    "status": "needs_review",
    "label": "Needs Review",
    "method": "unverified",
    "verified_at": null
  },
  "uncertainty": [
    "The work concerns individual animal identification. The material does not report specific results for other retrieval domains or establish that the protocol transfers unchanged."
  ],
  "connected_context": {
    "meaning": "CFR makes shortcut use directly testable by comparing identification on full images, subject-only foregrounds, and shortcut-only views. It complements visual stress and corruption evaluations with a controlled way to check whether image retrieval depends on the subject or surrounding cues. The evidence is for individual animal identification; transfer to other retrieval tasks remains unestablished.",
    "corpus_size": 1009,
    "generated_at": "2026-11-05T11:04:31.498Z",
    "connections": [
      {
        "title": "SABRE: Scalable and Automated Benchmarking of VLMs under Stress",
        "source_name": "arXiv",
        "source_url": "https://arxiv.org/abs/2608.07435v1",
        "feed7_url": "https://feed7.dev/p/2608-07435v1-0h6gzdk",
        "reason": "SABRE constructs visual stress tests, while CFR provides controlled subject and shortcut views; together they cover generated adversarial cases and tests of which image content supports a result."
      },
      {
        "title": "The Alignment Illusion in Multimodal Large Language Models",
        "source_name": "arXiv",
        "source_url": "https://arxiv.org/abs/2609.30210v1",
        "feed7_url": "https://feed7.dev/p/2609-30210v1-198in2y",
        "reason": "The alignment study tests performance after visual corruption; CFR tests performance when subject or shortcut content is isolated, giving complementary controls for whether visual information drives a result."
      }
    ]
  },
  "lifecycle": "Current",
  "published_at": "2026-11-04T18:19:26.000Z",
  "modified_at": "2026-11-04T18:19:26.000Z",
  "supersedes": [],
  "expires_at": null,
  "formats": {
    "html": "https://feed7.dev/p/2611-06864v1-1huxo8d",
    "json": "https://feed7.dev/p/2611-06864v1-1huxo8d.json",
    "markdown": "https://feed7.dev/p/2611-06864v1-1huxo8d.md"
  }
}