{ "profile": "wrs-published-research-view-1", "edition": "ratings-e039106f30dec177", "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS", "entry": { "id": "dabney-et-al-implicit-quantile-network-2018", "title": "Implicit Quantile Networks: sampled quantile-function value models", "summary": "Conditioned a neural value model on sampled quantile fractions to approximate return quantile functions, supporting distributional learning and distorted-quantile risk-sensitive action evaluation.", "boundary": "Extends QR-DQN rather than inventing distributional RL or quantile regression. Finite-capacity stochastic quantile sampling is not exact representation of arbitrary return laws; risk-distortion experiments do not establish universally optimal risk-sensitive control or neural convergence.", "area": "Distributional reinforcement learning", "topic": "Artificial intelligence", "year": 2018, "lane": "HUMAN", "result": "ALGORITHM", "package_id": "dabney-et-al-implicit-quantile-network-2018", "target": { "record_id": "urn:uuid:eff2cfd0-d87c-4828-a1b9-07df64af24cf", "version": 1, "content_hash": "sha256:1514aa9f5f4a3122cb4de186e5e77eebaccc173db34f70cdc7027af4569cd142" }, "humans": [ "Will Dabney", "Georg Ostrovski", "David Silver", "Rémi Munos" ], "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": "cur20261015-arxiv-1806-06923v1", "title": "Implicit Quantile Networks for Distributional Reinforcement Learning", "url": "https://arxiv.org/abs/1806.06923v1" } ], "points": "125000", "published_grade": { "id": "dabney-et-al-implicit-quantile-network-2018", "entry_id": "dabney-et-al-implicit-quantile-network-2018", "band": "A", "modifier": null, "target": { "record_id": "urn:uuid:eff2cfd0-d87c-4828-a1b9-07df64af24cf", "version": 1, "content_hash": "sha256:1514aa9f5f4a3122cb4de186e5e77eebaccc173db34f70cdc7027af4569cd142" }, "what_changed": "Conditioned a neural value model on sampled quantile fractions to approximate return quantile functions, supporting distributional learning and distorted-quantile risk-sensitive action evaluation.", "scope": "Conditioned a neural value model on sampled quantile fractions to approximate return quantile functions, supporting distributional learning and distorted-quantile risk-sensitive action evaluation. Extends QR-DQN rather than inventing distributional RL or quantile regression. Finite-capacity stochastic quantile sampling is not exact representation of arbitrary return laws; risk-distortion experiments do not establish universally optimal risk-sensitive control or neural convergence.", "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:eff2cfd0-d87c-4828-a1b9-07df64af24cf", "version": 1, "content_hash": "sha256:1514aa9f5f4a3122cb4de186e5e77eebaccc173db34f70cdc7027af4569cd142" }, "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": "Will Dabney", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Implicit Quantile Networks for Distributional Reinforcement Learning", "url": "https://arxiv.org/abs/1806.06923v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Georg Ostrovski", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Implicit Quantile Networks for Distributional Reinforcement Learning", "url": "https://arxiv.org/abs/1806.06923v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "David Silver", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Implicit Quantile Networks for Distributional Reinforcement Learning", "url": "https://arxiv.org/abs/1806.06923v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Rémi Munos", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "Implicit Quantile Networks for Distributional Reinforcement Learning", "url": "https://arxiv.org/abs/1806.06923v1" } } ], "evidence": [ { "description": "Bibliographic/stated-result source; not a WRS proof verification", "id": "source-0", "kind": "source_reference", "source": { "citation": "Implicit Quantile Networks for Distributional Reinforcement Learning", "url": "https://arxiv.org/abs/1806.06923v1" } } ], "origin": "HUMAN_LEGACY", "record_profile": "wrs-local-research-notes-0.1", "record_type": "research_note", "rrp_version": "0.1", "scope": { "description": "Extends QR-DQN rather than inventing distributional RL or quantile regression. Finite-capacity stochastic quantile sampling is not exact representation of arbitrary return laws; risk-distortion experiments do not establish universally optimal risk-sensitive control or neural convergence.", "domain": "Distributional reinforcement learning" }, "sources": [ { "citation": "Implicit Quantile Networks for Distributional Reinforcement Learning", "url": "https://arxiv.org/abs/1806.06923v1" } ], "statement": "# Implicit Quantile Networks: sampled quantile-function value models\n\nConditioned a neural value model on sampled quantile fractions to approximate return quantile functions, supporting distributional learning and distorted-quantile risk-sensitive action evaluation.\n\n## Scientific scope\n\nExtends QR-DQN rather than inventing distributional RL or quantile regression. Finite-capacity stochastic quantile sampling is not exact representation of arbitrary return laws; risk-distortion experiments do not establish universally optimal risk-sensitive control or neural convergence.\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\nWill Dabney; Georg Ostrovski; David Silver; Rémi Munos\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- [Implicit Quantile Networks for Distributional Reinforcement Learning](https://arxiv.org/abs/1806.06923v1)\n", "summary": "Conditioned a neural value model on sampled quantile fractions to approximate return quantile functions, supporting distributional learning and distorted-quantile risk-sensitive action evaluation.", "title": "Implicit Quantile Networks: sampled quantile-function value models" }, "record_id": "urn:uuid:eff2cfd0-d87c-4828-a1b9-07df64af24cf", "version": 1, "source_ref": { "record_id": "urn:uuid:eff2cfd0-d87c-4828-a1b9-07df64af24cf", "version": 1, "content_hash": "sha256:1514aa9f5f4a3122cb4de186e5e77eebaccc173db34f70cdc7027af4569cd142" }, "export_view": "wrs-public-editorial-view-1", "public_view_hash": "sha256:8a176988899fc2b4b6d655b37a0ef5b172e31f70ccbc844c945c697ad27a21c7" }, "target": { "record_id": "urn:uuid:eff2cfd0-d87c-4828-a1b9-07df64af24cf", "version": 1, "content_hash": "sha256:1514aa9f5f4a3122cb4de186e5e77eebaccc173db34f70cdc7027af4569cd142" }, "published_grade": { "profile": "wrs-published-grade-1", "edition": "2026-10-28.1", "id": "dabney-et-al-implicit-quantile-network-2018", "entry_id": "dabney-et-al-implicit-quantile-network-2018", "revision": 1, "title": "Implicit Quantile Networks: sampled quantile-function value models", "context": "Distributional reinforcement learning", "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": "Conditioned a neural value model on sampled quantile fractions to approximate return quantile functions, supporting distributional learning and distorted-quantile risk-sensitive action evaluation.", "scope": "Conditioned a neural value model on sampled quantile fractions to approximate return quantile functions, supporting distributional learning and distorted-quantile risk-sensitive action evaluation. Extends QR-DQN rather than inventing distributional RL or quantile regression. Finite-capacity stochastic quantile sampling is not exact representation of arbitrary return laws; risk-distortion experiments do not establish universally optimal risk-sensitive control or neural convergence.", "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. Extends QR-DQN rather than inventing distributional RL or quantile regression. Finite-capacity stochastic quantile sampling is not exact representation of arbitrary return laws; risk-distortion experiments do not establish universally optimal risk-sensitive control or neural convergence.", "references": [ { "title": "Implicit Quantile Networks for Distributional Reinforcement Learning", "url": "https://arxiv.org/abs/1806.06923v1", "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:eff2cfd0-d87c-4828-a1b9-07df64af24cf", "version": 1, "content_hash": "sha256:1514aa9f5f4a3122cb4de186e5e77eebaccc173db34f70cdc7027af4569cd142" }, "scientific_correctness_certified": false, "independent_expert_calibration": false, "method": "PROPRIETARY", "points": "125000" }, "canonical_agent_submission": false, "scientific_verification_awarded": false, "reader_guide": null, "reading_status": "SOURCE_LINKED_NOTE" }