{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:YL2UTCIQZH2DETA37TS65VKZVB","short_pith_number":"pith:YL2UTCIQ","schema_version":"1.0","canonical_sha256":"c2f5498910c9f4324c1bfce5eed559a84860f86e6108f7c95dee30386050b7ae","source":{"kind":"arxiv","id":"2605.08253","version":2},"attestation_state":"computed","paper":{"title":"Path-Coupled Bellman Flows for Distributional Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Path-Coupled Bellman Flows learn return distributions by matching flows along source-consistent paths that couple current and successor distributions through shared noise.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Boyang Xu, Hao Yan, Qing Zou, Siqin Yang","submitted_at":"2026-05-07T19:05:01Z","abstract_excerpt":"Distributional reinforcement learning (DRL) models the full return distribution, but existing finite-support or quantile-based methods rely on projections, while recent flow-based approaches can suffer from \\emph{boundary mismatch} at the flow source or from \\emph{high-variance} bootstrapping when current and successor noises are independent. We propose Path-Coupled Bellman Flows (PCBF), a continuous-time DRL method that learns return distributions with flow matching using \\textbf{source-consistent Bellman-coupled paths}: the current path starts from the required base prior at $t{=}0$, reaches"},"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":"2605.08253","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-07T19:05:01Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8f1de45d5c5fa5cfca37a95d0668f5d2a280f260bae70ef254b4460868c5aead","abstract_canon_sha256":"ad05c7ad54970ab15a01c138808279e97e8b709b2b3d00f6d050b285446159b7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T00:13:46.841300Z","signature_b64":"JBY+ivk4clYB+KeH08cfiAPqB7pd1APHrNkIWq30fbQ84u8KdelHoY46+K25htNmtqwpIu44klxULxwSP9ZkDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c2f5498910c9f4324c1bfce5eed559a84860f86e6108f7c95dee30386050b7ae","last_reissued_at":"2026-06-05T00:13:46.840649Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T00:13:46.840649Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Path-Coupled Bellman Flows for Distributional Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Path-Coupled Bellman Flows learn return distributions by matching flows along source-consistent paths that couple current and successor distributions through shared noise.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Boyang Xu, Hao Yan, Qing Zou, Siqin Yang","submitted_at":"2026-05-07T19:05:01Z","abstract_excerpt":"Distributional reinforcement learning (DRL) models the full return distribution, but existing finite-support or quantile-based methods rely on projections, while recent flow-based approaches can suffer from \\emph{boundary mismatch} at the flow source or from \\emph{high-variance} bootstrapping when current and successor noises are independent. We propose Path-Coupled Bellman Flows (PCBF), a continuous-time DRL method that learns return distributions with flow matching using \\textbf{source-consistent Bellman-coupled paths}: the current path starts from the required base prior at $t{=}0$, reaches"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"PCBF learns return distributions with flow matching using source-consistent Bellman-coupled paths: the current path starts from the required base prior at t=0, reaches the Bellman target at t=1, and maintains a pathwise affine relation to the successor flow at intermediate times, with a lambda-parameterized control-variate target that trades controlled bias for variance reduction.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the pathwise affine relation and coupling via shared base noise preserve the necessary distributional properties for correct Bellman updates without requiring time-t marginals to satisfy a distributional Bellman fixed point for all t, and that this coupling does not introduce new inconsistencies in the continuous-time flow.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Path-Coupled Bellman Flows use source-consistent Bellman-coupled paths and a lambda-parameterized control-variate to learn return distributions via flow matching, improving fidelity and stability over prior DRL approaches.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Path-Coupled Bellman Flows learn return distributions by matching flows along source-consistent paths that couple current and successor distributions through shared noise.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"c999209a3a003948ddfefab264abec217fd4b1b4105355e3ddf829567aa73d63"},"source":{"id":"2605.08253","kind":"arxiv","version":2},"verdict":{"id":"bfbb9445-33a8-44e6-a159-7e3da3345d02","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-12T02:03:35.025384Z","strongest_claim":"PCBF learns return distributions with flow matching using source-consistent Bellman-coupled paths: the current path starts from the required base prior at t=0, reaches the Bellman target at t=1, and maintains a pathwise affine relation to the successor flow at intermediate times, with a lambda-parameterized control-variate target that trades controlled bias for variance reduction.","one_line_summary":"Path-Coupled Bellman Flows use source-consistent Bellman-coupled paths and a lambda-parameterized control-variate to learn return distributions via flow matching, improving fidelity and stability over prior DRL approaches.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the pathwise affine relation and coupling via shared base noise preserve the necessary distributional properties for correct Bellman updates without requiring time-t marginals to satisfy a distributional Bellman fixed point for all t, and that this coupling does not introduce new inconsistencies in the continuous-time flow.","pith_extraction_headline":"Path-Coupled Bellman Flows learn return distributions by matching flows along source-consistent paths that couple current and successor distributions through shared noise."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.08253/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-20T12:02:03.721352Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-20T07:33:40.011882Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T17:31:19.623104Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T12:20:49.106433Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"a542f4b50cbea51e8d3eaa106e7723854b849f88c5e79bd802d127aaff35983c"},"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":"2605.08253","created_at":"2026-06-05T00:13:46.840740+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.08253v2","created_at":"2026-06-05T00:13:46.840740+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.08253","created_at":"2026-06-05T00:13:46.840740+00:00"},{"alias_kind":"pith_short_12","alias_value":"YL2UTCIQZH2D","created_at":"2026-06-05T00:13:46.840740+00:00"},{"alias_kind":"pith_short_16","alias_value":"YL2UTCIQZH2DETA3","created_at":"2026-06-05T00:13:46.840740+00:00"},{"alias_kind":"pith_short_8","alias_value":"YL2UTCIQ","created_at":"2026-06-05T00:13:46.840740+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/YL2UTCIQZH2DETA37TS65VKZVB","json":"https://pith.science/pith/YL2UTCIQZH2DETA37TS65VKZVB.json","graph_json":"https://pith.science/api/pith-number/YL2UTCIQZH2DETA37TS65VKZVB/graph.json","events_json":"https://pith.science/api/pith-number/YL2UTCIQZH2DETA37TS65VKZVB/events.json","paper":"https://pith.science/paper/YL2UTCIQ"},"agent_actions":{"view_html":"https://pith.science/pith/YL2UTCIQZH2DETA37TS65VKZVB","download_json":"https://pith.science/pith/YL2UTCIQZH2DETA37TS65VKZVB.json","view_paper":"https://pith.science/paper/YL2UTCIQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.08253&json=true","fetch_graph":"https://pith.science/api/pith-number/YL2UTCIQZH2DETA37TS65VKZVB/graph.json","fetch_events":"https://pith.science/api/pith-number/YL2UTCIQZH2DETA37TS65VKZVB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YL2UTCIQZH2DETA37TS65VKZVB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YL2UTCIQZH2DETA37TS65VKZVB/action/storage_attestation","attest_author":"https://pith.science/pith/YL2UTCIQZH2DETA37TS65VKZVB/action/author_attestation","sign_citation":"https://pith.science/pith/YL2UTCIQZH2DETA37TS65VKZVB/action/citation_signature","submit_replication":"https://pith.science/pith/YL2UTCIQZH2DETA37TS65VKZVB/action/replication_record"}},"created_at":"2026-06-05T00:13:46.840740+00:00","updated_at":"2026-06-05T00:13:46.840740+00:00"}