{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:F53AASVBTE574QXMNWH4QHGLTY","short_pith_number":"pith:F53AASVB","schema_version":"1.0","canonical_sha256":"2f76004aa1993bfe42ec6d8fc81ccb9e3b0d24da21ce64b2d12400037a137725","source":{"kind":"arxiv","id":"2311.02546","version":4},"attestation_state":"computed","paper":{"title":"On the Second-Order Convergence of Biased Policy Gradient Algorithms","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Diego Klabjan, Siqiao Mu","submitted_at":"2023-11-05T02:33:30Z","abstract_excerpt":"Since the objective functions of reinforcement learning problems are typically highly nonconvex, it is desirable that policy gradient, the most popular algorithm, escapes saddle points and arrives at second-order stationary points. Existing results only consider vanilla policy gradient algorithms with unbiased gradient estimators, but practical implementations under the infinite-horizon discounted reward setting are biased due to finite-horizon sampling. Moreover, actor-critic methods, whose second-order convergence has not yet been established, are also biased due to the critic approximation "},"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":"2311.02546","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-11-05T02:33:30Z","cross_cats_sorted":[],"title_canon_sha256":"13167ee8f38d9947e67c56ca9a40aa62531c02e5ed863eb553fbe979ffdab284","abstract_canon_sha256":"01cbfb06af85288048c5374213d937a75f6e88baaf696afcfb59bb1339dd2c54"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:18:45.970903Z","signature_b64":"cG64IVQMH/x7nlFRVuEpFYYtTi13DYmHSLo6DqUEij+89BUWjJFM24JkDRUk6AFyH9LkysdGNGyLRwTRGqOPDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2f76004aa1993bfe42ec6d8fc81ccb9e3b0d24da21ce64b2d12400037a137725","last_reissued_at":"2026-07-05T08:18:45.970395Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:18:45.970395Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On the Second-Order Convergence of Biased Policy Gradient Algorithms","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Diego Klabjan, Siqiao Mu","submitted_at":"2023-11-05T02:33:30Z","abstract_excerpt":"Since the objective functions of reinforcement learning problems are typically highly nonconvex, it is desirable that policy gradient, the most popular algorithm, escapes saddle points and arrives at second-order stationary points. Existing results only consider vanilla policy gradient algorithms with unbiased gradient estimators, but practical implementations under the infinite-horizon discounted reward setting are biased due to finite-horizon sampling. Moreover, actor-critic methods, whose second-order convergence has not yet been established, are also biased due to the critic approximation "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.02546","kind":"arxiv","version":4},"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/2311.02546/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":"2311.02546","created_at":"2026-07-05T08:18:45.970455+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.02546v4","created_at":"2026-07-05T08:18:45.970455+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.02546","created_at":"2026-07-05T08:18:45.970455+00:00"},{"alias_kind":"pith_short_12","alias_value":"F53AASVBTE57","created_at":"2026-07-05T08:18:45.970455+00:00"},{"alias_kind":"pith_short_16","alias_value":"F53AASVBTE574QXM","created_at":"2026-07-05T08:18:45.970455+00:00"},{"alias_kind":"pith_short_8","alias_value":"F53AASVB","created_at":"2026-07-05T08:18:45.970455+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/F53AASVBTE574QXMNWH4QHGLTY","json":"https://pith.science/pith/F53AASVBTE574QXMNWH4QHGLTY.json","graph_json":"https://pith.science/api/pith-number/F53AASVBTE574QXMNWH4QHGLTY/graph.json","events_json":"https://pith.science/api/pith-number/F53AASVBTE574QXMNWH4QHGLTY/events.json","paper":"https://pith.science/paper/F53AASVB"},"agent_actions":{"view_html":"https://pith.science/pith/F53AASVBTE574QXMNWH4QHGLTY","download_json":"https://pith.science/pith/F53AASVBTE574QXMNWH4QHGLTY.json","view_paper":"https://pith.science/paper/F53AASVB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.02546&json=true","fetch_graph":"https://pith.science/api/pith-number/F53AASVBTE574QXMNWH4QHGLTY/graph.json","fetch_events":"https://pith.science/api/pith-number/F53AASVBTE574QXMNWH4QHGLTY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/F53AASVBTE574QXMNWH4QHGLTY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/F53AASVBTE574QXMNWH4QHGLTY/action/storage_attestation","attest_author":"https://pith.science/pith/F53AASVBTE574QXMNWH4QHGLTY/action/author_attestation","sign_citation":"https://pith.science/pith/F53AASVBTE574QXMNWH4QHGLTY/action/citation_signature","submit_replication":"https://pith.science/pith/F53AASVBTE574QXMNWH4QHGLTY/action/replication_record"}},"created_at":"2026-07-05T08:18:45.970455+00:00","updated_at":"2026-07-05T08:18:45.970455+00:00"}