{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2016:TLDAE4I26GZVXLI4BKM6J22ZOT","short_pith_number":"pith:TLDAE4I2","schema_version":"1.0","canonical_sha256":"9ac602711af1b35bad1c0a99e4eb5974fe854b6bc6bea2fb9229dbc8f8c6af51","source":{"kind":"arxiv","id":"1602.02658","version":4},"attestation_state":"computed","paper":{"title":"Graying the black box: Understanding DQNs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.NE"],"primary_cat":"cs.LG","authors_text":"Nir Ben Zrihem, Shie Mannor, Tom Zahavy","submitted_at":"2016-02-08T17:27:31Z","abstract_excerpt":"In recent years there is a growing interest in using deep representations for reinforcement learning. In this paper, we present a methodology and tools to analyze Deep Q-networks (DQNs) in a non-blind matter. Moreover, we propose a new model, the Semi Aggregated Markov Decision Process (SAMDP), and an algorithm that learns it automatically. The SAMDP model allows us to identify spatio-temporal abstractions directly from features and may be used as a sub-goal detector in future work. Using our tools we reveal that the features learned by DQNs aggregate the state space in a hierarchical fashion,"},"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":"1602.02658","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2016-02-08T17:27:31Z","cross_cats_sorted":["cs.AI","cs.NE"],"title_canon_sha256":"dd741a41fc2b8b400a93de5941ea837bfaf67c6638d3a4ebbece8bbed418aebf","abstract_canon_sha256":"a1fb8f9bcc34f333598bcb9d195605ab4dbba8614dd5bb11bc3eb212b9cd8136"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:45:54.493612Z","signature_b64":"AeefYWRYjMVRYm77c2prDvdamY+R7Jr0mZAdhhPYrg9UvwEI+okC0Q8f3RLuv8DkcwGUHWhZ6rIwJPdyBid3BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9ac602711af1b35bad1c0a99e4eb5974fe854b6bc6bea2fb9229dbc8f8c6af51","last_reissued_at":"2026-05-18T00:45:54.493014Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:45:54.493014Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Graying the black box: Understanding DQNs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.NE"],"primary_cat":"cs.LG","authors_text":"Nir Ben Zrihem, Shie Mannor, Tom Zahavy","submitted_at":"2016-02-08T17:27:31Z","abstract_excerpt":"In recent years there is a growing interest in using deep representations for reinforcement learning. In this paper, we present a methodology and tools to analyze Deep Q-networks (DQNs) in a non-blind matter. Moreover, we propose a new model, the Semi Aggregated Markov Decision Process (SAMDP), and an algorithm that learns it automatically. The SAMDP model allows us to identify spatio-temporal abstractions directly from features and may be used as a sub-goal detector in future work. Using our tools we reveal that the features learned by DQNs aggregate the state space in a hierarchical fashion,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1602.02658","kind":"arxiv","version":4},"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":"1602.02658","created_at":"2026-05-18T00:45:54.493094+00:00"},{"alias_kind":"arxiv_version","alias_value":"1602.02658v4","created_at":"2026-05-18T00:45:54.493094+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1602.02658","created_at":"2026-05-18T00:45:54.493094+00:00"},{"alias_kind":"pith_short_12","alias_value":"TLDAE4I26GZV","created_at":"2026-05-18T12:30:44.179134+00:00"},{"alias_kind":"pith_short_16","alias_value":"TLDAE4I26GZVXLI4","created_at":"2026-05-18T12:30:44.179134+00:00"},{"alias_kind":"pith_short_8","alias_value":"TLDAE4I2","created_at":"2026-05-18T12:30:44.179134+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/TLDAE4I26GZVXLI4BKM6J22ZOT","json":"https://pith.science/pith/TLDAE4I26GZVXLI4BKM6J22ZOT.json","graph_json":"https://pith.science/api/pith-number/TLDAE4I26GZVXLI4BKM6J22ZOT/graph.json","events_json":"https://pith.science/api/pith-number/TLDAE4I26GZVXLI4BKM6J22ZOT/events.json","paper":"https://pith.science/paper/TLDAE4I2"},"agent_actions":{"view_html":"https://pith.science/pith/TLDAE4I26GZVXLI4BKM6J22ZOT","download_json":"https://pith.science/pith/TLDAE4I26GZVXLI4BKM6J22ZOT.json","view_paper":"https://pith.science/paper/TLDAE4I2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1602.02658&json=true","fetch_graph":"https://pith.science/api/pith-number/TLDAE4I26GZVXLI4BKM6J22ZOT/graph.json","fetch_events":"https://pith.science/api/pith-number/TLDAE4I26GZVXLI4BKM6J22ZOT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TLDAE4I26GZVXLI4BKM6J22ZOT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TLDAE4I26GZVXLI4BKM6J22ZOT/action/storage_attestation","attest_author":"https://pith.science/pith/TLDAE4I26GZVXLI4BKM6J22ZOT/action/author_attestation","sign_citation":"https://pith.science/pith/TLDAE4I26GZVXLI4BKM6J22ZOT/action/citation_signature","submit_replication":"https://pith.science/pith/TLDAE4I26GZVXLI4BKM6J22ZOT/action/replication_record"}},"created_at":"2026-05-18T00:45:54.493094+00:00","updated_at":"2026-05-18T00:45:54.493094+00:00"}