{
  "profile": "wrs-published-research-view-1",
  "edition": "ratings-e039106f30dec177",
  "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS",
  "entry": {
    "id": "ai-oct-outlier-exposure",
    "title": "Outlier Exposure: anomaly detection trained with auxiliary outlier data",
    "summary": "Trains a representation or detector with a complementary outlier loss on a diverse auxiliary dataset disjoint from the test anomalies, improving selected image and text anomaly detection.",
    "boundary": "Generalization to unseen anomalies is empirically demonstrated, not universal. Choice and overlap of auxiliary data matter; original methodology differs from methods tuned to each tested OOD source and results are not automatically comparable.",
    "area": "AI uncertainty, reliability and selective prediction",
    "topic": "Artificial intelligence",
    "year": 2018,
    "lane": "HUMAN",
    "result": "ALGORITHM",
    "package_id": "ai-oct-outlier-exposure",
    "target": {
      "record_id": "urn:uuid:64f6dc52-92fb-4e65-a35d-d6867bd9bbf1",
      "version": 1,
      "content_hash": "sha256:c217e4d11def6a0ebf2a87185f29c2e5e4df42e1978f4552530fe31bdebdba3c"
    },
    "humans": [
      "Dan Hendrycks",
      "Mantas Mazeika",
      "Thomas G. Dietterich"
    ],
    "authors_complete": true,
    "human_role_note": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.",
    "agents": [],
    "related": [],
    "subresults": [],
    "scientific_assessment": {
      "source_review": "SOURCE_CURATED",
      "scientific_review": "NOT_REVIEWED_BY_WRS",
      "formal_replay": "NOT_RUN",
      "provider_attestation": "NONE"
    },
    "sources": [
      {
        "id": "src-ai-oct-outlier-exposure-1812-04606v1",
        "title": "Deep Anomaly Detection with Outlier Exposure",
        "url": "https://arxiv.org/pdf/1812.04606v1"
      }
    ],
    "points": "140000",
    "published_grade": {
      "id": "ai-oct-outlier-exposure",
      "entry_id": "ai-oct-outlier-exposure",
      "band": "A",
      "modifier": null,
      "target": {
        "record_id": "urn:uuid:64f6dc52-92fb-4e65-a35d-d6867bd9bbf1",
        "version": 1,
        "content_hash": "sha256:c217e4d11def6a0ebf2a87185f29c2e5e4df42e1978f4552530fe31bdebdba3c"
      },
      "what_changed": "Trains a representation or detector with a complementary outlier loss on a diverse auxiliary dataset disjoint from the test anomalies, improving selected image and text anomaly detection.",
      "scope": "Trains a representation or detector with a complementary outlier loss on a diverse auxiliary dataset disjoint from the test anomalies, improving selected image and text anomaly detection. Generalization to unseen anomalies is empirically demonstrated, not universal. Choice and overlap of auxiliary data matter; original methodology differs from methods tuned to each tested OOD source and results are not automatically comparable.",
      "status": "PROVISIONAL_EDITORIAL",
      "scientific_correctness_certified": false
    },
    "assessment_disposition": {
      "status": "ASSESSED",
      "label": "Provisionally assessed",
      "reason": "A current grade and separate reference score identify this exact contribution.",
      "target": {
        "record_id": "urn:uuid:64f6dc52-92fb-4e65-a35d-d6867bd9bbf1",
        "version": 1,
        "content_hash": "sha256:c217e4d11def6a0ebf2a87185f29c2e5e4df42e1978f4552530fe31bdebdba3c"
      },
      "edition": "ratings-e039106f30dec177"
    }
  },
  "version": {
    "kind": "PUBLIC_RECORD_VIEW",
    "previous": null,
    "published_at": "2026-10-16T00:00:00Z",
    "publisher": {
      "principal_id": "urn:uuid:80116931-785b-40a2-a24b-9d59db0c7e33",
      "role": "WRS_EDITORIAL_PUBLICATION"
    },
    "record": {
      "contributors": [
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Dan Hendrycks",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Deep Anomaly Detection with Outlier Exposure",
            "url": "https://arxiv.org/pdf/1812.04606v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Mantas Mazeika",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Deep Anomaly Detection with Outlier Exposure",
            "url": "https://arxiv.org/pdf/1812.04606v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Thomas G. Dietterich",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Deep Anomaly Detection with Outlier Exposure",
            "url": "https://arxiv.org/pdf/1812.04606v1"
          }
        }
      ],
      "evidence": [
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-0",
          "kind": "source_reference",
          "source": {
            "citation": "Deep Anomaly Detection with Outlier Exposure",
            "url": "https://arxiv.org/pdf/1812.04606v1"
          }
        }
      ],
      "origin": "HUMAN_LEGACY",
      "record_profile": "wrs-local-research-notes-0.1",
      "record_type": "research_note",
      "rrp_version": "0.1",
      "scope": {
        "description": "Generalization to unseen anomalies is empirically demonstrated, not universal. Choice and overlap of auxiliary data matter; original methodology differs from methods tuned to each tested OOD source and results are not automatically comparable.",
        "domain": "AI uncertainty, reliability and selective prediction"
      },
      "sources": [
        {
          "citation": "Deep Anomaly Detection with Outlier Exposure",
          "url": "https://arxiv.org/pdf/1812.04606v1"
        }
      ],
      "statement": "# Outlier Exposure: anomaly detection trained with auxiliary outlier data\n\nTrains a representation or detector with a complementary outlier loss on a diverse auxiliary dataset disjoint from the test anomalies, improving selected image and text anomaly detection.\n\n## Scientific scope\n\nGeneralization to unseen anomalies is empirically demonstrated, not universal. Choice and overlap of auxiliary data matter; original methodology differs from methods tuned to each tested OOD source and results are not automatically comparable.\n\n## Reading status\n\nThis is a source-linked WRS catalogue note. A full explanatory guide and independent scientific reconstruction have not been completed for this entry. Read the identified original work for its complete narrative and arguments.\n\n## People and systems\n\nDan Hendrycks; Mantas Mazeika; Thomas G. Dietterich\n\nSource-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.\n\n\n\n## Original sources\n\n- [Deep Anomaly Detection with Outlier Exposure](https://arxiv.org/pdf/1812.04606v1)\n",
      "summary": "Trains a representation or detector with a complementary outlier loss on a diverse auxiliary dataset disjoint from the test anomalies, improving selected image and text anomaly detection.",
      "title": "Outlier Exposure: anomaly detection trained with auxiliary outlier data"
    },
    "record_id": "urn:uuid:64f6dc52-92fb-4e65-a35d-d6867bd9bbf1",
    "version": 1,
    "source_ref": {
      "record_id": "urn:uuid:64f6dc52-92fb-4e65-a35d-d6867bd9bbf1",
      "version": 1,
      "content_hash": "sha256:c217e4d11def6a0ebf2a87185f29c2e5e4df42e1978f4552530fe31bdebdba3c"
    },
    "export_view": "wrs-public-editorial-view-1",
    "public_view_hash": "sha256:22299e4051ff5ccc4ae8bf152869619af99a78aefaf41fdb22659eda65fa1b61"
  },
  "target": {
    "record_id": "urn:uuid:64f6dc52-92fb-4e65-a35d-d6867bd9bbf1",
    "version": 1,
    "content_hash": "sha256:c217e4d11def6a0ebf2a87185f29c2e5e4df42e1978f4552530fe31bdebdba3c"
  },
  "published_grade": {
    "profile": "wrs-published-grade-1",
    "edition": "2026-10-28.1",
    "id": "ai-oct-outlier-exposure",
    "entry_id": "ai-oct-outlier-exposure",
    "revision": 1,
    "title": "Outlier Exposure: anomaly detection trained with auxiliary outlier data",
    "context": "AI uncertainty, reliability and selective prediction",
    "track": "Human-designed research method or resource",
    "band": "A",
    "modifier": null,
    "status": "PROVISIONAL_EDITORIAL",
    "assessed_on": "2026-10-16",
    "evidence_cutoff": "2026-10-16",
    "achievement_state": "SUPPORTED_SCOPED_RESULT",
    "what_changed": "Trains a representation or detector with a complementary outlier loss on a diverse auxiliary dataset disjoint from the test anomalies, improving selected image and text anomaly detection.",
    "scope": "Trains a representation or detector with a complementary outlier loss on a diverse auxiliary dataset disjoint from the test anomalies, improving selected image and text anomaly detection. Generalization to unseen anomalies is empirically demonstrated, not universal. Choice and overlap of auxiliary data matter; original methodology differs from methods tuned to each tested OOD source and results are not automatically comparable.",
    "evidence_summary": "The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Generalization to unseen anomalies is empirically demonstrated, not universal. Choice and overlap of auxiliary data matter; original methodology differs from methods tuned to each tested OOD source and results are not automatically comparable.",
    "references": [
      {
        "title": "Deep Anomaly Detection with Outlier Exposure",
        "url": "https://arxiv.org/pdf/1812.04606v1",
        "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification."
      }
    ],
    "roles": [
      {
        "who": "Source-listed authors and technical contributors",
        "credit": "SCIENCE",
        "did": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities."
      }
    ],
    "target": {
      "record_id": "urn:uuid:64f6dc52-92fb-4e65-a35d-d6867bd9bbf1",
      "version": 1,
      "content_hash": "sha256:c217e4d11def6a0ebf2a87185f29c2e5e4df42e1978f4552530fe31bdebdba3c"
    },
    "scientific_correctness_certified": false,
    "independent_expert_calibration": false,
    "method": "PROPRIETARY",
    "points": "140000"
  },
  "canonical_agent_submission": false,
  "scientific_verification_awarded": false,
  "reader_guide": null,
  "reading_status": "SOURCE_LINKED_NOTE"
}