{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2017:LRF3K7ZFVPACCTV2PZPTHY2W3D","short_pith_number":"pith:LRF3K7ZF","schema_version":"1.0","canonical_sha256":"5c4bb57f25abc0214eba7e5f33e356d8ef9b8ce2135bbc917f0986dc8a9c38e4","source":{"kind":"arxiv","id":"1702.02284","version":1},"attestation_state":"computed","paper":{"title":"Adversarial Attacks on Neural Network Policies","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","stat.ML"],"primary_cat":"cs.LG","authors_text":"Ian Goodfellow, Nicolas Papernot, Pieter Abbeel, Sandy Huang, Yan Duan","submitted_at":"2017-02-08T04:33:55Z","abstract_excerpt":"Machine learning classifiers are known to be vulnerable to inputs maliciously constructed by adversaries to force misclassification. Such adversarial examples have been extensively studied in the context of computer vision applications. In this work, we show adversarial attacks are also effective when targeting neural network policies in reinforcement learning. Specifically, we show existing adversarial example crafting techniques can be used to significantly degrade test-time performance of trained policies. Our threat model considers adversaries capable of introducing small perturbations to "},"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":"1702.02284","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-02-08T04:33:55Z","cross_cats_sorted":["cs.CR","stat.ML"],"title_canon_sha256":"e2e2ae06fb1ec29c45acc4f0ba2f6117f331eb1eda0552f8cc96fc209608fa4e","abstract_canon_sha256":"2ad7cf7a05e277008b4f03a8c0d36d7af058f6effce6714a3582bcf4d4f02409"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:51:06.296871Z","signature_b64":"PGHBU0OkuS91uhA8cbGk3N/3P4ltBFw9gWWQ5bcRoZJRAl2E7srlwsowIm/7RN15RnBcTPYNPPiqtHyu/h7JCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5c4bb57f25abc0214eba7e5f33e356d8ef9b8ce2135bbc917f0986dc8a9c38e4","last_reissued_at":"2026-05-18T00:51:06.296291Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:51:06.296291Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adversarial Attacks on Neural Network Policies","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","stat.ML"],"primary_cat":"cs.LG","authors_text":"Ian Goodfellow, Nicolas Papernot, Pieter Abbeel, Sandy Huang, Yan Duan","submitted_at":"2017-02-08T04:33:55Z","abstract_excerpt":"Machine learning classifiers are known to be vulnerable to inputs maliciously constructed by adversaries to force misclassification. Such adversarial examples have been extensively studied in the context of computer vision applications. In this work, we show adversarial attacks are also effective when targeting neural network policies in reinforcement learning. Specifically, we show existing adversarial example crafting techniques can be used to significantly degrade test-time performance of trained policies. Our threat model considers adversaries capable of introducing small perturbations to "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1702.02284","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":""},"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":"1702.02284","created_at":"2026-05-18T00:51:06.296378+00:00"},{"alias_kind":"arxiv_version","alias_value":"1702.02284v1","created_at":"2026-05-18T00:51:06.296378+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1702.02284","created_at":"2026-05-18T00:51:06.296378+00:00"},{"alias_kind":"pith_short_12","alias_value":"LRF3K7ZFVPAC","created_at":"2026-05-18T12:31:28.150371+00:00"},{"alias_kind":"pith_short_16","alias_value":"LRF3K7ZFVPACCTV2","created_at":"2026-05-18T12:31:28.150371+00:00"},{"alias_kind":"pith_short_8","alias_value":"LRF3K7ZF","created_at":"2026-05-18T12:31:28.150371+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":11,"internal_anchor_count":10,"sample":[{"citing_arxiv_id":"1906.12061","citing_title":"Learning to Cope with Adversarial Attacks","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2406.09250","citing_title":"MirrorCheck: Efficient Adversarial Defense for Vision-Language Models","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2502.03698","citing_title":"How Vulnerable Is My Learned Policy? Universal Adversarial Perturbation Attacks On Modern Behavior Cloning Policies","ref_index":28,"is_internal_anchor":true},{"citing_arxiv_id":"2502.02844","citing_title":"Wolfpack Adversarial Attack for Robust Multi-Agent Reinforcement Learning","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2510.01479","citing_title":"Density-Ratio Weighted Behavioral Cloning: Learning Control Policies from Corrupted Datasets","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2605.16312","citing_title":"When Actions Disappear: Adversarial Action Removal in Self-Play Reinforcement Learning","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2605.18058","citing_title":"Threats to Arabic Handwriting Recognition: Investigating Black-Box Adversarial Attacks on embedded ConvNet models","ref_index":23,"is_internal_anchor":true},{"citing_arxiv_id":"2512.09893","citing_title":"A Speculative GLRT-Backed ApproachRobust Deep Learning-Based Array Processing","ref_index":39,"is_internal_anchor":true},{"citing_arxiv_id":"2603.28281","citing_title":"Corruption-robust Offline Multi-agent Reinforcement Learning From Human Feedback","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12792","citing_title":"SoK: A Comprehensive Analysis of the Current Status of Neural Tangent Generalization Attacks with Research Directions","ref_index":42,"is_internal_anchor":true},{"citing_arxiv_id":"2605.02495","citing_title":"Efficient Preference Poisoning Attack on Offline RLHF","ref_index":124,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LRF3K7ZFVPACCTV2PZPTHY2W3D","json":"https://pith.science/pith/LRF3K7ZFVPACCTV2PZPTHY2W3D.json","graph_json":"https://pith.science/api/pith-number/LRF3K7ZFVPACCTV2PZPTHY2W3D/graph.json","events_json":"https://pith.science/api/pith-number/LRF3K7ZFVPACCTV2PZPTHY2W3D/events.json","paper":"https://pith.science/paper/LRF3K7ZF"},"agent_actions":{"view_html":"https://pith.science/pith/LRF3K7ZFVPACCTV2PZPTHY2W3D","download_json":"https://pith.science/pith/LRF3K7ZFVPACCTV2PZPTHY2W3D.json","view_paper":"https://pith.science/paper/LRF3K7ZF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1702.02284&json=true","fetch_graph":"https://pith.science/api/pith-number/LRF3K7ZFVPACCTV2PZPTHY2W3D/graph.json","fetch_events":"https://pith.science/api/pith-number/LRF3K7ZFVPACCTV2PZPTHY2W3D/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LRF3K7ZFVPACCTV2PZPTHY2W3D/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LRF3K7ZFVPACCTV2PZPTHY2W3D/action/storage_attestation","attest_author":"https://pith.science/pith/LRF3K7ZFVPACCTV2PZPTHY2W3D/action/author_attestation","sign_citation":"https://pith.science/pith/LRF3K7ZFVPACCTV2PZPTHY2W3D/action/citation_signature","submit_replication":"https://pith.science/pith/LRF3K7ZFVPACCTV2PZPTHY2W3D/action/replication_record"}},"created_at":"2026-05-18T00:51:06.296378+00:00","updated_at":"2026-05-18T00:51:06.296378+00:00"}