{ "profile": "wrs-published-research-view-1", "edition": "ratings-e039106f30dec177", "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS", "entry": { "id": "zhang-et-al-dhen-2022", "title": "DHEN: hierarchical ensembles of heterogeneous feature interactions", "summary": "Stacked ensembles of complementary feature-interaction modules with shared feature processing and residual information flow, paired with hybrid sharded training to support deeper click-prediction models.", "boundary": "Reuses established attention, convolution, DLRM, DCN, FSDP and checkpointing components. Its contribution is their heterogeneous interaction hierarchy and associated training design, not invention of those components. Normalized-entropy improvement is not a revenue or absolute click-through-rate gain; results are empirical.", "area": "Recommendation architectures", "topic": "Artificial intelligence", "year": 2022, "lane": "HUMAN", "result": "ALGORITHM", "package_id": "zhang-et-al-dhen-2022", "target": { "record_id": "urn:uuid:f79cedd5-0933-4e7b-aedb-54a5a34e59ab", "version": 1, "content_hash": "sha256:237cc1d8ac8d9adf9833b23e268358c1e078992518aff551bfb26e8ad35329e2" }, "humans": [ "Buyun Zhang", "Liang Luo", "Xi Liu", "Jay Li", "Zeliang Chen", "Weilin Zhang", "Xiaohan Wei", "Yuchen Hao", "Michael Tsang", "Wenjun Wang", "Yang Liu", "Huayu Li", "Yasmine Badr", "Jongsoo Park", "Jiyan Yang", "Dheevatsa Mudigere", "Ellie Wen" ], "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-2203-11014v1", "title": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } ], "points": "65000", "published_grade": { "id": "zhang-et-al-dhen-2022", "entry_id": "zhang-et-al-dhen-2022", "band": "B", "modifier": "+", "target": { "record_id": "urn:uuid:f79cedd5-0933-4e7b-aedb-54a5a34e59ab", "version": 1, "content_hash": "sha256:237cc1d8ac8d9adf9833b23e268358c1e078992518aff551bfb26e8ad35329e2" }, "what_changed": "Stacked ensembles of complementary feature-interaction modules with shared feature processing and residual information flow, paired with hybrid sharded training to support deeper click-prediction models.", "scope": "Stacked ensembles of complementary feature-interaction modules with shared feature processing and residual information flow, paired with hybrid sharded training to support deeper click-prediction models. Reuses established attention, convolution, DLRM, DCN, FSDP and checkpointing components. Its contribution is their heterogeneous interaction hierarchy and associated training design, not invention of those components. Normalized-entropy improvement is not a revenue or absolute click-through-rate gain; results are empirical.", "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:f79cedd5-0933-4e7b-aedb-54a5a34e59ab", "version": 1, "content_hash": "sha256:237cc1d8ac8d9adf9833b23e268358c1e078992518aff551bfb26e8ad35329e2" }, "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": "Buyun Zhang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Liang Luo", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Xi Liu", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Jay Li", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Zeliang Chen", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Weilin Zhang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Xiaohan Wei", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Yuchen Hao", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Michael Tsang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Wenjun Wang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Yang Liu", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Huayu Li", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Yasmine Badr", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Jongsoo Park", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Jiyan Yang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Dheevatsa Mudigere", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Ellie Wen", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } } ], "evidence": [ { "description": "Bibliographic/stated-result source; not a WRS proof verification", "id": "source-0", "kind": "source_reference", "source": { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } } ], "origin": "HUMAN_LEGACY", "record_profile": "wrs-local-research-notes-0.1", "record_type": "research_note", "rrp_version": "0.1", "scope": { "description": "Reuses established attention, convolution, DLRM, DCN, FSDP and checkpointing components. Its contribution is their heterogeneous interaction hierarchy and associated training design, not invention of those components. Normalized-entropy improvement is not a revenue or absolute click-through-rate gain; results are empirical.", "domain": "Recommendation architectures" }, "sources": [ { "citation": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1" } ], "statement": "# DHEN: hierarchical ensembles of heterogeneous feature interactions\n\nStacked ensembles of complementary feature-interaction modules with shared feature processing and residual information flow, paired with hybrid sharded training to support deeper click-prediction models.\n\n## Scientific scope\n\nReuses established attention, convolution, DLRM, DCN, FSDP and checkpointing components. Its contribution is their heterogeneous interaction hierarchy and associated training design, not invention of those components. Normalized-entropy improvement is not a revenue or absolute click-through-rate gain; results are empirical.\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\nBuyun Zhang; Liang Luo; Xi Liu; Jay Li; Zeliang Chen; Weilin Zhang; Xiaohan Wei; Yuchen Hao; Michael Tsang; Wenjun Wang; Yang Liu; Huayu Li; Yasmine Badr; Jongsoo Park; Jiyan Yang; Dheevatsa Mudigere; Ellie Wen\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- [DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction](https://arxiv.org/abs/2203.11014v1)\n", "summary": "Stacked ensembles of complementary feature-interaction modules with shared feature processing and residual information flow, paired with hybrid sharded training to support deeper click-prediction models.", "title": "DHEN: hierarchical ensembles of heterogeneous feature interactions" }, "record_id": "urn:uuid:f79cedd5-0933-4e7b-aedb-54a5a34e59ab", "version": 1, "source_ref": { "record_id": "urn:uuid:f79cedd5-0933-4e7b-aedb-54a5a34e59ab", "version": 1, "content_hash": "sha256:237cc1d8ac8d9adf9833b23e268358c1e078992518aff551bfb26e8ad35329e2" }, "export_view": "wrs-public-editorial-view-1", "public_view_hash": "sha256:a04e305d5d19f8a3856cc7ab70009defca1b054a5b0b7e135bb25d5996577fa9" }, "target": { "record_id": "urn:uuid:f79cedd5-0933-4e7b-aedb-54a5a34e59ab", "version": 1, "content_hash": "sha256:237cc1d8ac8d9adf9833b23e268358c1e078992518aff551bfb26e8ad35329e2" }, "published_grade": { "profile": "wrs-published-grade-1", "edition": "2026-10-28.1", "id": "zhang-et-al-dhen-2022", "entry_id": "zhang-et-al-dhen-2022", "revision": 1, "title": "DHEN: hierarchical ensembles of heterogeneous feature interactions", "context": "Recommendation architectures", "track": "Human-designed research method or resource", "band": "B", "modifier": "+", "status": "PROVISIONAL_EDITORIAL", "assessed_on": "2026-10-16", "evidence_cutoff": "2026-10-16", "achievement_state": "SUPPORTED_SCOPED_RESULT", "what_changed": "Stacked ensembles of complementary feature-interaction modules with shared feature processing and residual information flow, paired with hybrid sharded training to support deeper click-prediction models.", "scope": "Stacked ensembles of complementary feature-interaction modules with shared feature processing and residual information flow, paired with hybrid sharded training to support deeper click-prediction models. Reuses established attention, convolution, DLRM, DCN, FSDP and checkpointing components. Its contribution is their heterogeneous interaction hierarchy and associated training design, not invention of those components. Normalized-entropy improvement is not a revenue or absolute click-through-rate gain; results are empirical.", "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. Reuses established attention, convolution, DLRM, DCN, FSDP and checkpointing components. Its contribution is their heterogeneous interaction hierarchy and associated training design, not invention of those components. Normalized-entropy improvement is not a revenue or absolute click-through-rate gain; results are empirical.", "references": [ { "title": "DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction", "url": "https://arxiv.org/abs/2203.11014v1", "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:f79cedd5-0933-4e7b-aedb-54a5a34e59ab", "version": 1, "content_hash": "sha256:237cc1d8ac8d9adf9833b23e268358c1e078992518aff551bfb26e8ad35329e2" }, "scientific_correctness_certified": false, "independent_expert_calibration": false, "method": "PROPRIETARY", "points": "65000" }, "canonical_agent_submission": false, "scientific_verification_awarded": false, "reader_guide": null, "reading_status": "SOURCE_LINKED_NOTE" }