{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:KEOGMCYFDDTSVN3HCYNTHX4OMP","short_pith_number":"pith:KEOGMCYF","schema_version":"1.0","canonical_sha256":"511c660b0518e72ab767161b33df8e63de51581d738aefd407887da7a881c13b","source":{"kind":"arxiv","id":"2008.01976","version":2},"attestation_state":"computed","paper":{"title":"Robust Deep Reinforcement Learning through Adversarial Loss","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR","stat.ML"],"primary_cat":"cs.LG","authors_text":"Alexandre Megretski, Luca Daniel, Tsui-Wei Weng, Tuomas Oikarinen, Wang Zhang","submitted_at":"2020-08-05T07:49:42Z","abstract_excerpt":"Recent studies have shown that deep reinforcement learning agents are vulnerable to small adversarial perturbations on the agent's inputs, which raises concerns about deploying such agents in the real world. To address this issue, we propose RADIAL-RL, a principled framework to train reinforcement learning agents with improved robustness against $l_p$-norm bounded adversarial attacks. Our framework is compatible with popular deep reinforcement learning algorithms and we demonstrate its performance with deep Q-learning, A3C and PPO. We experiment on three deep RL benchmarks (Atari, MuJoCo and P"},"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":"2008.01976","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-08-05T07:49:42Z","cross_cats_sorted":["cs.AI","cs.CR","stat.ML"],"title_canon_sha256":"021c899ad7341ac24f2d47eb0012f11d8e46cda66aef61c34e0bad0c8545b3de","abstract_canon_sha256":"f0803efa43b08246cb1080fbd81f2aec86df374b4fee4e061564e1299c4f5070"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:30:50.046200Z","signature_b64":"fVSO3G/TjcQmWWx/XcDHT81wj6g5II2+31Orb/1LQvbl9VVFaCQzRMg4+HwymsVBQ0uDRrnjOygG44RqZd8ABw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"511c660b0518e72ab767161b33df8e63de51581d738aefd407887da7a881c13b","last_reissued_at":"2026-07-05T03:30:50.045777Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:30:50.045777Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Robust Deep Reinforcement Learning through Adversarial Loss","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR","stat.ML"],"primary_cat":"cs.LG","authors_text":"Alexandre Megretski, Luca Daniel, Tsui-Wei Weng, Tuomas Oikarinen, Wang Zhang","submitted_at":"2020-08-05T07:49:42Z","abstract_excerpt":"Recent studies have shown that deep reinforcement learning agents are vulnerable to small adversarial perturbations on the agent's inputs, which raises concerns about deploying such agents in the real world. To address this issue, we propose RADIAL-RL, a principled framework to train reinforcement learning agents with improved robustness against $l_p$-norm bounded adversarial attacks. Our framework is compatible with popular deep reinforcement learning algorithms and we demonstrate its performance with deep Q-learning, A3C and PPO. We experiment on three deep RL benchmarks (Atari, MuJoCo and P"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2008.01976","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/2008.01976/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":"2008.01976","created_at":"2026-07-05T03:30:50.045836+00:00"},{"alias_kind":"arxiv_version","alias_value":"2008.01976v2","created_at":"2026-07-05T03:30:50.045836+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2008.01976","created_at":"2026-07-05T03:30:50.045836+00:00"},{"alias_kind":"pith_short_12","alias_value":"KEOGMCYFDDTS","created_at":"2026-07-05T03:30:50.045836+00:00"},{"alias_kind":"pith_short_16","alias_value":"KEOGMCYFDDTSVN3H","created_at":"2026-07-05T03:30:50.045836+00:00"},{"alias_kind":"pith_short_8","alias_value":"KEOGMCYF","created_at":"2026-07-05T03:30:50.045836+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/KEOGMCYFDDTSVN3HCYNTHX4OMP","json":"https://pith.science/pith/KEOGMCYFDDTSVN3HCYNTHX4OMP.json","graph_json":"https://pith.science/api/pith-number/KEOGMCYFDDTSVN3HCYNTHX4OMP/graph.json","events_json":"https://pith.science/api/pith-number/KEOGMCYFDDTSVN3HCYNTHX4OMP/events.json","paper":"https://pith.science/paper/KEOGMCYF"},"agent_actions":{"view_html":"https://pith.science/pith/KEOGMCYFDDTSVN3HCYNTHX4OMP","download_json":"https://pith.science/pith/KEOGMCYFDDTSVN3HCYNTHX4OMP.json","view_paper":"https://pith.science/paper/KEOGMCYF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2008.01976&json=true","fetch_graph":"https://pith.science/api/pith-number/KEOGMCYFDDTSVN3HCYNTHX4OMP/graph.json","fetch_events":"https://pith.science/api/pith-number/KEOGMCYFDDTSVN3HCYNTHX4OMP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KEOGMCYFDDTSVN3HCYNTHX4OMP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KEOGMCYFDDTSVN3HCYNTHX4OMP/action/storage_attestation","attest_author":"https://pith.science/pith/KEOGMCYFDDTSVN3HCYNTHX4OMP/action/author_attestation","sign_citation":"https://pith.science/pith/KEOGMCYFDDTSVN3HCYNTHX4OMP/action/citation_signature","submit_replication":"https://pith.science/pith/KEOGMCYFDDTSVN3HCYNTHX4OMP/action/replication_record"}},"created_at":"2026-07-05T03:30:50.045836+00:00","updated_at":"2026-07-05T03:30:50.045836+00:00"}