{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2020:O7NYUHPLXJTVMI7TBNJJ5W6PCM","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":"d820a315c1c45db24058e268b78e00d31c1bba552449ce45e61b6d092941f9f6","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-11-17T18:48:59Z","title_canon_sha256":"225c1f014a72a7dbd7a41935b042a433a308bc64e42504d2783ddd2290165de1"},"schema_version":"1.0","source":{"id":"2011.08827","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2011.08827","created_at":"2026-07-05T01:52:31Z"},{"alias_kind":"arxiv_version","alias_value":"2011.08827v1","created_at":"2026-07-05T01:52:31Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2011.08827","created_at":"2026-07-05T01:52:31Z"},{"alias_kind":"pith_short_12","alias_value":"O7NYUHPLXJTV","created_at":"2026-07-05T01:52:31Z"},{"alias_kind":"pith_short_16","alias_value":"O7NYUHPLXJTVMI7T","created_at":"2026-07-05T01:52:31Z"},{"alias_kind":"pith_short_8","alias_value":"O7NYUHPL","created_at":"2026-07-05T01:52:31Z"}],"graph_snapshots":[{"event_id":"sha256:a7f30d5c38bc20d4cd2e9f7a5aca438acc92ce69b4a32f6c876cb7d7d79a2c8c","target":"graph","created_at":"2026-07-05T01:52:31Z","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/2011.08827/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"How can we design agents that pursue a given objective when all feedback mechanisms are influenceable by the agent? Standard RL algorithms assume a secure reward function, and can thus perform poorly in settings where agents can tamper with the reward-generating mechanism. We present a principled solution to the problem of learning from influenceable feedback, which combines approval with a decoupled feedback collection procedure. For a natural class of corruption functions, decoupled approval algorithms have aligned incentives both at convergence and for their local updates. Empirically, they","authors_text":"Jonathan Uesato, Ramana Kumar, Richard Ngo, Shane Legg, Tom Everitt, Victoria Krakovna","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-11-17T18:48:59Z","title":"Avoiding Tampering Incentives in Deep RL via Decoupled Approval"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2011.08827","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":"sha256:7b5a7523c206c2cb504ab980596a1cb2eebdaae708d9bb1c3841f45dfbd5f515","target":"record","created_at":"2026-07-05T01:52:31Z","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":"d820a315c1c45db24058e268b78e00d31c1bba552449ce45e61b6d092941f9f6","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-11-17T18:48:59Z","title_canon_sha256":"225c1f014a72a7dbd7a41935b042a433a308bc64e42504d2783ddd2290165de1"},"schema_version":"1.0","source":{"id":"2011.08827","kind":"arxiv","version":1}},"canonical_sha256":"77db8a1debba675623f30b529edbcf1327bc2d83d9872467fd6a4a542d96bc56","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"77db8a1debba675623f30b529edbcf1327bc2d83d9872467fd6a4a542d96bc56","first_computed_at":"2026-07-05T01:52:31.371709Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T01:52:31.371709Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"xVEFen1kaz2e0sxEoxIiTqgyIdYCq6ZdYnXnDpJZcYy0jfjdJhBj2OFQxPQPulRfGGjjkDXj7T+DxpoEoX7YCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T01:52:31.372236Z","signed_message":"canonical_sha256_bytes"},"source_id":"2011.08827","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7b5a7523c206c2cb504ab980596a1cb2eebdaae708d9bb1c3841f45dfbd5f515","sha256:a7f30d5c38bc20d4cd2e9f7a5aca438acc92ce69b4a32f6c876cb7d7d79a2c8c"],"state_sha256":"52fc5b6c188a50ada558fa30a7c6a70a6031ffdff549f2dece4b5096224b086d"}