{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:I3VLMFEIC4TIDAJRPHPOPYXGAV","short_pith_number":"pith:I3VLMFEI","schema_version":"1.0","canonical_sha256":"46eab61488172681813179dee7e2e6054a2f19a0be5dbd659931ccd32f2acad5","source":{"kind":"arxiv","id":"2310.15418","version":1},"attestation_state":"computed","paper":{"title":"Fractal Landscapes in Policy Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Sicun Gao, Sylvia Herbert, Tao Wang","submitted_at":"2023-10-24T00:22:19Z","abstract_excerpt":"Policy gradient lies at the core of deep reinforcement learning (RL) in continuous domains. Despite much success, it is often observed in practice that RL training with policy gradient can fail for many reasons, even on standard control problems with known solutions. We propose a framework for understanding one inherent limitation of the policy gradient approach: the optimization landscape in the policy space can be extremely non-smooth or fractal for certain classes of MDPs, such that there does not exist gradient to be estimated in the first place. We draw on techniques from chaos theory and"},"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":"2310.15418","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-10-24T00:22:19Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ffc311654196fc93f78f0dc4c19f6b0d6483717e85e86085f0ef9308a685be11","abstract_canon_sha256":"cfffe22fd2c3c7aa338f189ba0cb113def12efa7b3688387c2bb35868e952eb4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:35:40.562852Z","signature_b64":"XGieC6mDI5kIBGO2ZZyT+PwJopgIegsIPGDOkkE3rxFsOz766moc112+Cn46DZxwEJxPRVNSq+m2Ww0tBzqsDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"46eab61488172681813179dee7e2e6054a2f19a0be5dbd659931ccd32f2acad5","last_reissued_at":"2026-07-05T07:35:40.562519Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:35:40.562519Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Fractal Landscapes in Policy Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Sicun Gao, Sylvia Herbert, Tao Wang","submitted_at":"2023-10-24T00:22:19Z","abstract_excerpt":"Policy gradient lies at the core of deep reinforcement learning (RL) in continuous domains. Despite much success, it is often observed in practice that RL training with policy gradient can fail for many reasons, even on standard control problems with known solutions. We propose a framework for understanding one inherent limitation of the policy gradient approach: the optimization landscape in the policy space can be extremely non-smooth or fractal for certain classes of MDPs, such that there does not exist gradient to be estimated in the first place. We draw on techniques from chaos theory and"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.15418","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/2310.15418/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":"2310.15418","created_at":"2026-07-05T07:35:40.562572+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.15418v1","created_at":"2026-07-05T07:35:40.562572+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.15418","created_at":"2026-07-05T07:35:40.562572+00:00"},{"alias_kind":"pith_short_12","alias_value":"I3VLMFEIC4TI","created_at":"2026-07-05T07:35:40.562572+00:00"},{"alias_kind":"pith_short_16","alias_value":"I3VLMFEIC4TIDAJR","created_at":"2026-07-05T07:35:40.562572+00:00"},{"alias_kind":"pith_short_8","alias_value":"I3VLMFEI","created_at":"2026-07-05T07:35:40.562572+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/I3VLMFEIC4TIDAJRPHPOPYXGAV","json":"https://pith.science/pith/I3VLMFEIC4TIDAJRPHPOPYXGAV.json","graph_json":"https://pith.science/api/pith-number/I3VLMFEIC4TIDAJRPHPOPYXGAV/graph.json","events_json":"https://pith.science/api/pith-number/I3VLMFEIC4TIDAJRPHPOPYXGAV/events.json","paper":"https://pith.science/paper/I3VLMFEI"},"agent_actions":{"view_html":"https://pith.science/pith/I3VLMFEIC4TIDAJRPHPOPYXGAV","download_json":"https://pith.science/pith/I3VLMFEIC4TIDAJRPHPOPYXGAV.json","view_paper":"https://pith.science/paper/I3VLMFEI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.15418&json=true","fetch_graph":"https://pith.science/api/pith-number/I3VLMFEIC4TIDAJRPHPOPYXGAV/graph.json","fetch_events":"https://pith.science/api/pith-number/I3VLMFEIC4TIDAJRPHPOPYXGAV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/I3VLMFEIC4TIDAJRPHPOPYXGAV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/I3VLMFEIC4TIDAJRPHPOPYXGAV/action/storage_attestation","attest_author":"https://pith.science/pith/I3VLMFEIC4TIDAJRPHPOPYXGAV/action/author_attestation","sign_citation":"https://pith.science/pith/I3VLMFEIC4TIDAJRPHPOPYXGAV/action/citation_signature","submit_replication":"https://pith.science/pith/I3VLMFEIC4TIDAJRPHPOPYXGAV/action/replication_record"}},"created_at":"2026-07-05T07:35:40.562572+00:00","updated_at":"2026-07-05T07:35:40.562572+00:00"}