{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2019:YFCCBSQE2VLRA7E6OP65O2UZ3T","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":"cd6ba1b33431b30dfbd860563d9af206c8d0a0e3c8f18403b6261c969e255796","cross_cats_sorted":["cs.RO","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-10-09T00:45:53Z","title_canon_sha256":"611eff132183e03af5ce898a90b03e728badb303a55ed00f3f328f8701d7766a"},"schema_version":"1.0","source":{"id":"1910.03732","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1910.03732","created_at":"2026-07-05T00:10:57Z"},{"alias_kind":"arxiv_version","alias_value":"1910.03732v1","created_at":"2026-07-05T00:10:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1910.03732","created_at":"2026-07-05T00:10:57Z"},{"alias_kind":"pith_short_12","alias_value":"YFCCBSQE2VLR","created_at":"2026-07-05T00:10:57Z"},{"alias_kind":"pith_short_16","alias_value":"YFCCBSQE2VLRA7E6","created_at":"2026-07-05T00:10:57Z"},{"alias_kind":"pith_short_8","alias_value":"YFCCBSQE","created_at":"2026-07-05T00:10:57Z"}],"graph_snapshots":[{"event_id":"sha256:b56832321fca4d42eac5c599300b7033f1d60e65a36b66ebe510730f53f1af0a","target":"graph","created_at":"2026-07-05T00:10:57Z","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/1910.03732/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"When learning behavior, training data is often generated by the learner itself; this can result in unstable training dynamics, and this problem has particularly important applications in safety-sensitive real-world control tasks such as robotics. In this work, we propose a principled and model-agnostic approach to mitigate the issue of unstable learning dynamics by maintaining a history of a reinforcement learning agent over the course of training, and reverting to the parameters of a previous agent whenever performance significantly decreases. We develop techniques for evaluating this perform","authors_text":"Jake Bruce, J\\\"urgen Leitner, Thierry Peynot, Vibhavari Dasagi","cross_cats":["cs.RO","stat.ML"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-10-09T00:45:53Z","title":"Ctrl-Z: Recovering from Instability in Reinforcement Learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1910.03732","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":"sha256:8d169c15cf867c7ae17b5b978fac7ea2d4f81b43d78e9f2cf25ebe8f724af726","target":"record","created_at":"2026-07-05T00:10:57Z","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":"cd6ba1b33431b30dfbd860563d9af206c8d0a0e3c8f18403b6261c969e255796","cross_cats_sorted":["cs.RO","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-10-09T00:45:53Z","title_canon_sha256":"611eff132183e03af5ce898a90b03e728badb303a55ed00f3f328f8701d7766a"},"schema_version":"1.0","source":{"id":"1910.03732","kind":"arxiv","version":1}},"canonical_sha256":"c14420ca04d557107c9e73fdd76a99dcc4fb83995f6149e595f6377defd4c1fb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c14420ca04d557107c9e73fdd76a99dcc4fb83995f6149e595f6377defd4c1fb","first_computed_at":"2026-07-05T00:10:57.375766Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T00:10:57.375766Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"rsFrd7eKow3hHsDXYalODJEzhql6x4vsy8H27e+LdbQHA6auFbC6sBhh1EbjaGNz7Y7Gwx9Q2IgiyQC8/a/aCA==","signature_status":"signed_v1","signed_at":"2026-07-05T00:10:57.376210Z","signed_message":"canonical_sha256_bytes"},"source_id":"1910.03732","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8d169c15cf867c7ae17b5b978fac7ea2d4f81b43d78e9f2cf25ebe8f724af726","sha256:b56832321fca4d42eac5c599300b7033f1d60e65a36b66ebe510730f53f1af0a"],"state_sha256":"2a005c72e6f5f3c97412366c8877e0847f471a1ff8690226b123363d8bdcebef"}