{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2021:LNUPY4HRBGRTCOW6PRRXVYZQ5D","short_pith_number":"pith:LNUPY4HR","canonical_record":{"source":{"id":"2102.05371","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-02-10T10:27:49Z","cross_cats_sorted":[],"title_canon_sha256":"92d90fb23e8f9ffec20da8bd06ce92401c1cee34a6ad82f778080ed5720c851f","abstract_canon_sha256":"24c9110a3d15bac6969b7883386bd1a72966d3c0044be8578d9e2aefe06568a2"},"schema_version":"1.0"},"canonical_sha256":"5b68fc70f109a3313ade7c637ae330e8d1773893715c7f3cdc228aee051e7708","source":{"kind":"arxiv","id":"2102.05371","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2102.05371","created_at":"2026-07-05T02:14:17Z"},{"alias_kind":"arxiv_version","alias_value":"2102.05371v1","created_at":"2026-07-05T02:14:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.05371","created_at":"2026-07-05T02:14:17Z"},{"alias_kind":"pith_short_12","alias_value":"LNUPY4HRBGRT","created_at":"2026-07-05T02:14:17Z"},{"alias_kind":"pith_short_16","alias_value":"LNUPY4HRBGRTCOW6","created_at":"2026-07-05T02:14:17Z"},{"alias_kind":"pith_short_8","alias_value":"LNUPY4HR","created_at":"2026-07-05T02:14:17Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2021:LNUPY4HRBGRTCOW6PRRXVYZQ5D","target":"record","payload":{"canonical_record":{"source":{"id":"2102.05371","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-02-10T10:27:49Z","cross_cats_sorted":[],"title_canon_sha256":"92d90fb23e8f9ffec20da8bd06ce92401c1cee34a6ad82f778080ed5720c851f","abstract_canon_sha256":"24c9110a3d15bac6969b7883386bd1a72966d3c0044be8578d9e2aefe06568a2"},"schema_version":"1.0"},"canonical_sha256":"5b68fc70f109a3313ade7c637ae330e8d1773893715c7f3cdc228aee051e7708","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:14:17.984993Z","signature_b64":"vTogqxqnl4/O5Vautk08oypt6cVBE7Z3Uu9+kJZeTnWlnpIFVO9vb4EibSdeAut4O7YAKWD7a9+v6pWYhXe7Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5b68fc70f109a3313ade7c637ae330e8d1773893715c7f3cdc228aee051e7708","last_reissued_at":"2026-07-05T02:14:17.984558Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:14:17.984558Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2102.05371","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T02:14:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dHOb8mf0p7sGzwcTjGlkfLYGxDT07O71t5p30UNmptGmZFehWjVJKm5L3ZS5tfeLEV1QPqKrQtf5WhUuOOxKCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T22:55:36.558193Z"},"content_sha256":"1769e3e4b4b9ea0bfddf628885e4f8beb1efa3e018ef7cf0145865b7adb77d7e","schema_version":"1.0","event_id":"sha256:1769e3e4b4b9ea0bfddf628885e4f8beb1efa3e018ef7cf0145865b7adb77d7e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2021:LNUPY4HRBGRTCOW6PRRXVYZQ5D","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Risk-Averse Offline Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Andreas Krause, N\\'uria Armengol Urp\\'i, Sebastian Curi","submitted_at":"2021-02-10T10:27:49Z","abstract_excerpt":"Training Reinforcement Learning (RL) agents in high-stakes applications might be too prohibitive due to the risk associated to exploration. Thus, the agent can only use data previously collected by safe policies. While previous work considers optimizing the average performance using offline data, we focus on optimizing a risk-averse criteria, namely the CVaR. In particular, we present the Offline Risk-Averse Actor-Critic (O-RAAC), a model-free RL algorithm that is able to learn risk-averse policies in a fully offline setting. We show that O-RAAC learns policies with higher CVaR than risk-neutr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.05371","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/2102.05371/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":"2026-07-05T02:14:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Bt5u8MB0XF+mD8JEW3ltBUHQtQgUuWv3c5Ubwd5IUKmfTNow4XMIYBCBGZUE68W9+aUUqewPfzy4FbXLTFlNCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T22:55:36.558776Z"},"content_sha256":"175f4edbe5aad4f81b86e083ceba6bba242ad07e00bcfd253a9a8bb8baaf6a44","schema_version":"1.0","event_id":"sha256:175f4edbe5aad4f81b86e083ceba6bba242ad07e00bcfd253a9a8bb8baaf6a44"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LNUPY4HRBGRTCOW6PRRXVYZQ5D/bundle.json","state_url":"https://pith.science/pith/LNUPY4HRBGRTCOW6PRRXVYZQ5D/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LNUPY4HRBGRTCOW6PRRXVYZQ5D/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":"2026-08-06T22:55:36Z","links":{"resolver":"https://pith.science/pith/LNUPY4HRBGRTCOW6PRRXVYZQ5D","bundle":"https://pith.science/pith/LNUPY4HRBGRTCOW6PRRXVYZQ5D/bundle.json","state":"https://pith.science/pith/LNUPY4HRBGRTCOW6PRRXVYZQ5D/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LNUPY4HRBGRTCOW6PRRXVYZQ5D/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2021:LNUPY4HRBGRTCOW6PRRXVYZQ5D","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":"24c9110a3d15bac6969b7883386bd1a72966d3c0044be8578d9e2aefe06568a2","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-02-10T10:27:49Z","title_canon_sha256":"92d90fb23e8f9ffec20da8bd06ce92401c1cee34a6ad82f778080ed5720c851f"},"schema_version":"1.0","source":{"id":"2102.05371","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2102.05371","created_at":"2026-07-05T02:14:17Z"},{"alias_kind":"arxiv_version","alias_value":"2102.05371v1","created_at":"2026-07-05T02:14:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.05371","created_at":"2026-07-05T02:14:17Z"},{"alias_kind":"pith_short_12","alias_value":"LNUPY4HRBGRT","created_at":"2026-07-05T02:14:17Z"},{"alias_kind":"pith_short_16","alias_value":"LNUPY4HRBGRTCOW6","created_at":"2026-07-05T02:14:17Z"},{"alias_kind":"pith_short_8","alias_value":"LNUPY4HR","created_at":"2026-07-05T02:14:17Z"}],"graph_snapshots":[{"event_id":"sha256:175f4edbe5aad4f81b86e083ceba6bba242ad07e00bcfd253a9a8bb8baaf6a44","target":"graph","created_at":"2026-07-05T02:14:17Z","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/2102.05371/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Training Reinforcement Learning (RL) agents in high-stakes applications might be too prohibitive due to the risk associated to exploration. Thus, the agent can only use data previously collected by safe policies. While previous work considers optimizing the average performance using offline data, we focus on optimizing a risk-averse criteria, namely the CVaR. In particular, we present the Offline Risk-Averse Actor-Critic (O-RAAC), a model-free RL algorithm that is able to learn risk-averse policies in a fully offline setting. We show that O-RAAC learns policies with higher CVaR than risk-neutr","authors_text":"Andreas Krause, N\\'uria Armengol Urp\\'i, Sebastian Curi","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-02-10T10:27:49Z","title":"Risk-Averse Offline Reinforcement Learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.05371","kind":"arxiv","version":1},"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":"sha256:1769e3e4b4b9ea0bfddf628885e4f8beb1efa3e018ef7cf0145865b7adb77d7e","target":"record","created_at":"2026-07-05T02:14:17Z","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":"24c9110a3d15bac6969b7883386bd1a72966d3c0044be8578d9e2aefe06568a2","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-02-10T10:27:49Z","title_canon_sha256":"92d90fb23e8f9ffec20da8bd06ce92401c1cee34a6ad82f778080ed5720c851f"},"schema_version":"1.0","source":{"id":"2102.05371","kind":"arxiv","version":1}},"canonical_sha256":"5b68fc70f109a3313ade7c637ae330e8d1773893715c7f3cdc228aee051e7708","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5b68fc70f109a3313ade7c637ae330e8d1773893715c7f3cdc228aee051e7708","first_computed_at":"2026-07-05T02:14:17.984558Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T02:14:17.984558Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"vTogqxqnl4/O5Vautk08oypt6cVBE7Z3Uu9+kJZeTnWlnpIFVO9vb4EibSdeAut4O7YAKWD7a9+v6pWYhXe7Bg==","signature_status":"signed_v1","signed_at":"2026-07-05T02:14:17.984993Z","signed_message":"canonical_sha256_bytes"},"source_id":"2102.05371","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1769e3e4b4b9ea0bfddf628885e4f8beb1efa3e018ef7cf0145865b7adb77d7e","sha256:175f4edbe5aad4f81b86e083ceba6bba242ad07e00bcfd253a9a8bb8baaf6a44"],"state_sha256":"5d0d9fdbd5f0a0a263a82616ea45139e4895de0bdd983afc3ad940783c63a338"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dnlSMG1IpHWNGfr3Kx6h/tHcJxiAYi51ujjKjdReyO978EyIeOhVl5ItjNEM48rTiuBdlSh6sc7bByNdMC5YDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T22:55:36.563671Z","bundle_sha256":"596731ac235d603310b3cf1ee55de826363fd01a7e79646b08765ba2f4761323"}}