{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:FS33GFSGJWHULP2HUOUHPSRHFI","short_pith_number":"pith:FS33GFSG","schema_version":"1.0","canonical_sha256":"2cb7b316464d8f45bf47a3a877ca272a184401a982af03300a4adc2de25b5086","source":{"kind":"arxiv","id":"2608.02034","version":1},"attestation_state":"computed","paper":{"title":"Upper-Expectile Multi-Step Q-Learning for Off-Policy Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Abdelghani Ghanem, Mounir Ghogho","submitted_at":"2026-08-03T10:30:56Z","abstract_excerpt":"Multi-step returns accelerate reward propagation in off-policy reinforcement learning, but couple the evaluation of each decision to the suboptimal logged actions that follow it, inducing a pessimistic bias that grows with the horizon. We propose Expectile $n$-step Q-learning (ENQ), which replaces the symmetric $n$-step temporal-difference (TD) loss with an asymmetric expectile loss on the action-value error, with expectile level $\\tau$ as the only method-specific hyperparameter added beyond $n$-step TD. We prove that the ENQ operator is a $\\gamma^{n}$-contraction. Under deterministic dynamics"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2608.02034","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-08-03T10:30:56Z","cross_cats_sorted":[],"title_canon_sha256":"7669d41187197c84215673cd983553d7a015d5a135d79dea81c0a5d0d7cbb1ff","abstract_canon_sha256":"8e037adb53143c6ad7f76b176d6baced6465020869bd5287a1f2af565ea5719d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-04T02:10:32.904573Z","signature_b64":"UPHEnS81DNmMtDj3ORbQO2drakCP13Dkm7/BKNd6UtfLbYWvvyRhjH3WkMBsyw2KYIFYCYNPvDrGNR6B/VXWDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2cb7b316464d8f45bf47a3a877ca272a184401a982af03300a4adc2de25b5086","last_reissued_at":"2026-08-04T02:10:32.902633Z","signature_status":"signed_v1","first_computed_at":"2026-08-04T02:10:32.902633Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Upper-Expectile Multi-Step Q-Learning for Off-Policy Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Abdelghani Ghanem, Mounir Ghogho","submitted_at":"2026-08-03T10:30:56Z","abstract_excerpt":"Multi-step returns accelerate reward propagation in off-policy reinforcement learning, but couple the evaluation of each decision to the suboptimal logged actions that follow it, inducing a pessimistic bias that grows with the horizon. We propose Expectile $n$-step Q-learning (ENQ), which replaces the symmetric $n$-step temporal-difference (TD) loss with an asymmetric expectile loss on the action-value error, with expectile level $\\tau$ as the only method-specific hyperparameter added beyond $n$-step TD. We prove that the ENQ operator is a $\\gamma^{n}$-contraction. Under deterministic dynamics"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.02034","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2608.02034/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2608.02034","created_at":"2026-08-04T02:10:32.904496+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.02034v1","created_at":"2026-08-04T02:10:32.904496+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.02034","created_at":"2026-08-04T02:10:32.904496+00:00"},{"alias_kind":"pith_short_12","alias_value":"FS33GFSGJWHU","created_at":"2026-08-04T02:10:32.904496+00:00"},{"alias_kind":"pith_short_16","alias_value":"FS33GFSGJWHULP2H","created_at":"2026-08-04T02:10:32.904496+00:00"},{"alias_kind":"pith_short_8","alias_value":"FS33GFSG","created_at":"2026-08-04T02:10:32.904496+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FS33GFSGJWHULP2HUOUHPSRHFI","json":"https://pith.science/pith/FS33GFSGJWHULP2HUOUHPSRHFI.json","graph_json":"https://pith.science/api/pith-number/FS33GFSGJWHULP2HUOUHPSRHFI/graph.json","events_json":"https://pith.science/api/pith-number/FS33GFSGJWHULP2HUOUHPSRHFI/events.json","paper":"https://pith.science/paper/FS33GFSG"},"agent_actions":{"view_html":"https://pith.science/pith/FS33GFSGJWHULP2HUOUHPSRHFI","download_json":"https://pith.science/pith/FS33GFSGJWHULP2HUOUHPSRHFI.json","view_paper":"https://pith.science/paper/FS33GFSG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.02034&json=true","fetch_graph":"https://pith.science/api/pith-number/FS33GFSGJWHULP2HUOUHPSRHFI/graph.json","fetch_events":"https://pith.science/api/pith-number/FS33GFSGJWHULP2HUOUHPSRHFI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FS33GFSGJWHULP2HUOUHPSRHFI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FS33GFSGJWHULP2HUOUHPSRHFI/action/storage_attestation","attest_author":"https://pith.science/pith/FS33GFSGJWHULP2HUOUHPSRHFI/action/author_attestation","sign_citation":"https://pith.science/pith/FS33GFSGJWHULP2HUOUHPSRHFI/action/citation_signature","submit_replication":"https://pith.science/pith/FS33GFSGJWHULP2HUOUHPSRHFI/action/replication_record"}},"created_at":"2026-08-04T02:10:32.904496+00:00","updated_at":"2026-08-04T02:10:32.904496+00:00"}