{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2021:U7I65CTFDLNOAWM7RPVG4RVGR4","short_pith_number":"pith:U7I65CTF","canonical_record":{"source":{"id":"2106.04895","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-06-09T08:28:55Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"e0fdb56c50cbd59f942dc3f10789d1a47e1e4a557262f16d21c497bebe5195e5","abstract_canon_sha256":"ca0770498a062d1ff78ce81c39c2bd661edd8f382e799e63490a544efa424296"},"schema_version":"1.0"},"canonical_sha256":"a7d1ee8a651adae0599f8bea6e46a68f1b7e502e1ff2f0265a74d78b08d48911","source":{"kind":"arxiv","id":"2106.04895","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2106.04895","created_at":"2026-07-05T03:56:01Z"},{"alias_kind":"arxiv_version","alias_value":"2106.04895v2","created_at":"2026-07-05T03:56:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.04895","created_at":"2026-07-05T03:56:01Z"},{"alias_kind":"pith_short_12","alias_value":"U7I65CTFDLNO","created_at":"2026-07-05T03:56:01Z"},{"alias_kind":"pith_short_16","alias_value":"U7I65CTFDLNOAWM7","created_at":"2026-07-05T03:56:01Z"},{"alias_kind":"pith_short_8","alias_value":"U7I65CTF","created_at":"2026-07-05T03:56:01Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2021:U7I65CTFDLNOAWM7RPVG4RVGR4","target":"record","payload":{"canonical_record":{"source":{"id":"2106.04895","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-06-09T08:28:55Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"e0fdb56c50cbd59f942dc3f10789d1a47e1e4a557262f16d21c497bebe5195e5","abstract_canon_sha256":"ca0770498a062d1ff78ce81c39c2bd661edd8f382e799e63490a544efa424296"},"schema_version":"1.0"},"canonical_sha256":"a7d1ee8a651adae0599f8bea6e46a68f1b7e502e1ff2f0265a74d78b08d48911","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:56:01.896511Z","signature_b64":"sgQ6z8ra4X3beMI2dZmneqUEE4dNKbNX3ATCWqxmrfOnqdoTRX+jxEPyTkidkqEp1LQglJUaYv91nv1f+SH4DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a7d1ee8a651adae0599f8bea6e46a68f1b7e502e1ff2f0265a74d78b08d48911","last_reissued_at":"2026-07-05T03:56:01.896036Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:56:01.896036Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2106.04895","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T03:56:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"VtYmJDsvhWq8HMXHbJ58NrU8R6c/9/3gDpzPXoDK9yjZkZVC3PRU38WtnLFbtUf/psARiKu59x2gHig1QxQeCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T00:03:14.129844Z"},"content_sha256":"9f95d58d1ad8469593defad5ba939cb898d36b8f7a0192ea441491a004a9605d","schema_version":"1.0","event_id":"sha256:9f95d58d1ad8469593defad5ba939cb898d36b8f7a0192ea441491a004a9605d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2021:U7I65CTFDLNOAWM7RPVG4RVGR4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Policy Finetuning: Bridging Sample-Efficient Offline and Online Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Caiming Xiong, Huan Wang, Nan Jiang, Tengyang Xie, Yu Bai","submitted_at":"2021-06-09T08:28:55Z","abstract_excerpt":"Recent theoretical work studies sample-efficient reinforcement learning (RL) extensively in two settings: learning interactively in the environment (online RL), or learning from an offline dataset (offline RL). However, existing algorithms and theories for learning near-optimal policies in these two settings are rather different and disconnected. Towards bridging this gap, this paper initiates the theoretical study of policy finetuning, that is, online RL where the learner has additional access to a \"reference policy\" $\\mu$ close to the optimal policy $\\pi_\\star$ in a certain sense. We conside"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2106.04895","kind":"arxiv","version":2},"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/2106.04895/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T03:56:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"T35263bYoxe7rx+nsn7+8A7DMy4g3jIwF7T3YmJD8xKwDsbutnXVWUdvo/zQ/sS2U2gUwVPwPzsdagS6MAX+CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T00:03:14.131071Z"},"content_sha256":"2dd5a68add51a3c79014ac85f67843cc26bf7ec70eb71acbcf63d42d41a4e64a","schema_version":"1.0","event_id":"sha256:2dd5a68add51a3c79014ac85f67843cc26bf7ec70eb71acbcf63d42d41a4e64a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/U7I65CTFDLNOAWM7RPVG4RVGR4/bundle.json","state_url":"https://pith.science/pith/U7I65CTFDLNOAWM7RPVG4RVGR4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/U7I65CTFDLNOAWM7RPVG4RVGR4/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-11T00:03:14Z","links":{"resolver":"https://pith.science/pith/U7I65CTFDLNOAWM7RPVG4RVGR4","bundle":"https://pith.science/pith/U7I65CTFDLNOAWM7RPVG4RVGR4/bundle.json","state":"https://pith.science/pith/U7I65CTFDLNOAWM7RPVG4RVGR4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/U7I65CTFDLNOAWM7RPVG4RVGR4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2021:U7I65CTFDLNOAWM7RPVG4RVGR4","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"ca0770498a062d1ff78ce81c39c2bd661edd8f382e799e63490a544efa424296","cross_cats_sorted":["stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-06-09T08:28:55Z","title_canon_sha256":"e0fdb56c50cbd59f942dc3f10789d1a47e1e4a557262f16d21c497bebe5195e5"},"schema_version":"1.0","source":{"id":"2106.04895","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2106.04895","created_at":"2026-07-05T03:56:01Z"},{"alias_kind":"arxiv_version","alias_value":"2106.04895v2","created_at":"2026-07-05T03:56:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.04895","created_at":"2026-07-05T03:56:01Z"},{"alias_kind":"pith_short_12","alias_value":"U7I65CTFDLNO","created_at":"2026-07-05T03:56:01Z"},{"alias_kind":"pith_short_16","alias_value":"U7I65CTFDLNOAWM7","created_at":"2026-07-05T03:56:01Z"},{"alias_kind":"pith_short_8","alias_value":"U7I65CTF","created_at":"2026-07-05T03:56:01Z"}],"graph_snapshots":[{"event_id":"sha256:2dd5a68add51a3c79014ac85f67843cc26bf7ec70eb71acbcf63d42d41a4e64a","target":"graph","created_at":"2026-07-05T03:56:01Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2106.04895/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent theoretical work studies sample-efficient reinforcement learning (RL) extensively in two settings: learning interactively in the environment (online RL), or learning from an offline dataset (offline RL). However, existing algorithms and theories for learning near-optimal policies in these two settings are rather different and disconnected. Towards bridging this gap, this paper initiates the theoretical study of policy finetuning, that is, online RL where the learner has additional access to a \"reference policy\" $\\mu$ close to the optimal policy $\\pi_\\star$ in a certain sense. We conside","authors_text":"Caiming Xiong, Huan Wang, Nan Jiang, Tengyang Xie, Yu Bai","cross_cats":["stat.ML"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-06-09T08:28:55Z","title":"Policy Finetuning: Bridging Sample-Efficient Offline and Online Reinforcement Learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2106.04895","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:9f95d58d1ad8469593defad5ba939cb898d36b8f7a0192ea441491a004a9605d","target":"record","created_at":"2026-07-05T03:56:01Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"ca0770498a062d1ff78ce81c39c2bd661edd8f382e799e63490a544efa424296","cross_cats_sorted":["stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-06-09T08:28:55Z","title_canon_sha256":"e0fdb56c50cbd59f942dc3f10789d1a47e1e4a557262f16d21c497bebe5195e5"},"schema_version":"1.0","source":{"id":"2106.04895","kind":"arxiv","version":2}},"canonical_sha256":"a7d1ee8a651adae0599f8bea6e46a68f1b7e502e1ff2f0265a74d78b08d48911","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a7d1ee8a651adae0599f8bea6e46a68f1b7e502e1ff2f0265a74d78b08d48911","first_computed_at":"2026-07-05T03:56:01.896036Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T03:56:01.896036Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"sgQ6z8ra4X3beMI2dZmneqUEE4dNKbNX3ATCWqxmrfOnqdoTRX+jxEPyTkidkqEp1LQglJUaYv91nv1f+SH4DA==","signature_status":"signed_v1","signed_at":"2026-07-05T03:56:01.896511Z","signed_message":"canonical_sha256_bytes"},"source_id":"2106.04895","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9f95d58d1ad8469593defad5ba939cb898d36b8f7a0192ea441491a004a9605d","sha256:2dd5a68add51a3c79014ac85f67843cc26bf7ec70eb71acbcf63d42d41a4e64a"],"state_sha256":"8e97fea7e4b0c44a3b46a7df6bdecdcf1ed135fcd54bbe3ea41de46462df36b1"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"YCUr2BIOGJRK8cMz8Jwexzs+t4iuGWbNsHVQRXNjQ5Sxv5wfG75e53BVbNzEyhUNX+rQ4i4L5kSdgClXi8A+Aw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-11T00:03:14.137508Z","bundle_sha256":"2364f880f17dfa75c953293e201b6817a28a264225a25df7f7c9b173439d2499"}}