{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2017:UUS5F4LZZRQZWTFMJA4X3C53VW","short_pith_number":"pith:UUS5F4LZ","canonical_record":{"source":{"id":"1711.00138","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2017-10-31T23:03:17Z","cross_cats_sorted":[],"title_canon_sha256":"9ca38b59837320715bbb374620a7e5dc0fa7da050e2c3caba0003867b12351c9","abstract_canon_sha256":"161490d60f5586ae63dfe10db392c79a18bcd4292cab00dd9f480a7f7a0ffa0b"},"schema_version":"1.0"},"canonical_sha256":"a525d2f179cc619b4cac48397d8bbbadb266afac7c47a95230df4b622cbfd3ac","source":{"kind":"arxiv","id":"1711.00138","version":5},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1711.00138","created_at":"2026-05-18T00:06:05Z"},{"alias_kind":"arxiv_version","alias_value":"1711.00138v5","created_at":"2026-05-18T00:06:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1711.00138","created_at":"2026-05-18T00:06:05Z"},{"alias_kind":"pith_short_12","alias_value":"UUS5F4LZZRQZ","created_at":"2026-05-18T12:31:49Z"},{"alias_kind":"pith_short_16","alias_value":"UUS5F4LZZRQZWTFM","created_at":"2026-05-18T12:31:49Z"},{"alias_kind":"pith_short_8","alias_value":"UUS5F4LZ","created_at":"2026-05-18T12:31:49Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2017:UUS5F4LZZRQZWTFMJA4X3C53VW","target":"record","payload":{"canonical_record":{"source":{"id":"1711.00138","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2017-10-31T23:03:17Z","cross_cats_sorted":[],"title_canon_sha256":"9ca38b59837320715bbb374620a7e5dc0fa7da050e2c3caba0003867b12351c9","abstract_canon_sha256":"161490d60f5586ae63dfe10db392c79a18bcd4292cab00dd9f480a7f7a0ffa0b"},"schema_version":"1.0"},"canonical_sha256":"a525d2f179cc619b4cac48397d8bbbadb266afac7c47a95230df4b622cbfd3ac","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:06:05.810882Z","signature_b64":"RyKNi36iSK38TuiO7jaaZ7jrwljedfA9BkrDs3KpcLW5f8Jc2lYNCHTy34LsLzngCkaMzEAtF8q7VCN9UQcXAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a525d2f179cc619b4cac48397d8bbbadb266afac7c47a95230df4b622cbfd3ac","last_reissued_at":"2026-05-18T00:06:05.810269Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:06:05.810269Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1711.00138","source_version":5,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T00:06:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"U6yCxkP/qsZvy9+RDdgSbsAXAAcR0iTXJCkKi0Ze2gCArI7faY9waslmbvQs/BoKhO8VEfnJjfEWY+l1B46sBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T17:45:25.094432Z"},"content_sha256":"fa15f6fd696ef1c42d3bf8f906550a3b1f818c5972584c74f7e1537d55bceec9","schema_version":"1.0","event_id":"sha256:fa15f6fd696ef1c42d3bf8f906550a3b1f818c5972584c74f7e1537d55bceec9"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2017:UUS5F4LZZRQZWTFMJA4X3C53VW","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Visualizing and Understanding Atari Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Alan Fern, Anurag Koul, Jonathan Dodge, Sam Greydanus","submitted_at":"2017-10-31T23:03:17Z","abstract_excerpt":"While deep reinforcement learning (deep RL) agents are effective at maximizing rewards, it is often unclear what strategies they use to do so. In this paper, we take a step toward explaining deep RL agents through a case study using Atari 2600 environments. In particular, we focus on using saliency maps to understand how an agent learns and executes a policy. We introduce a method for generating useful saliency maps and use it to show 1) what strong agents attend to, 2) whether agents are making decisions for the right or wrong reasons, and 3) how agents evolve during learning. We also test ou"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1711.00138","kind":"arxiv","version":5},"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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T00:06:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1UAzRVWoi7b98D1+k9Rn65pcOyQfq2o6KYx2c27is6Br9GhrtKhg1ZQLNdIoi6P8KB3KWSAbDKZ1Mx2K2dE/Ag==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T17:45:25.095274Z"},"content_sha256":"1a996eab4dfaf3f7b0b3fd53d0524ad52ab5b278a48863e5cebbe6f3182ca1a2","schema_version":"1.0","event_id":"sha256:1a996eab4dfaf3f7b0b3fd53d0524ad52ab5b278a48863e5cebbe6f3182ca1a2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/UUS5F4LZZRQZWTFMJA4X3C53VW/bundle.json","state_url":"https://pith.science/pith/UUS5F4LZZRQZWTFMJA4X3C53VW/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/UUS5F4LZZRQZWTFMJA4X3C53VW/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-10T17:45:25Z","links":{"resolver":"https://pith.science/pith/UUS5F4LZZRQZWTFMJA4X3C53VW","bundle":"https://pith.science/pith/UUS5F4LZZRQZWTFMJA4X3C53VW/bundle.json","state":"https://pith.science/pith/UUS5F4LZZRQZWTFMJA4X3C53VW/state.json","well_known_bundle":"https://pith.science/.well-known/pith/UUS5F4LZZRQZWTFMJA4X3C53VW/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2017:UUS5F4LZZRQZWTFMJA4X3C53VW","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":"161490d60f5586ae63dfe10db392c79a18bcd4292cab00dd9f480a7f7a0ffa0b","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2017-10-31T23:03:17Z","title_canon_sha256":"9ca38b59837320715bbb374620a7e5dc0fa7da050e2c3caba0003867b12351c9"},"schema_version":"1.0","source":{"id":"1711.00138","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1711.00138","created_at":"2026-05-18T00:06:05Z"},{"alias_kind":"arxiv_version","alias_value":"1711.00138v5","created_at":"2026-05-18T00:06:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1711.00138","created_at":"2026-05-18T00:06:05Z"},{"alias_kind":"pith_short_12","alias_value":"UUS5F4LZZRQZ","created_at":"2026-05-18T12:31:49Z"},{"alias_kind":"pith_short_16","alias_value":"UUS5F4LZZRQZWTFM","created_at":"2026-05-18T12:31:49Z"},{"alias_kind":"pith_short_8","alias_value":"UUS5F4LZ","created_at":"2026-05-18T12:31:49Z"}],"graph_snapshots":[{"event_id":"sha256:1a996eab4dfaf3f7b0b3fd53d0524ad52ab5b278a48863e5cebbe6f3182ca1a2","target":"graph","created_at":"2026-05-18T00:06:05Z","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"},"paper":{"abstract_excerpt":"While deep reinforcement learning (deep RL) agents are effective at maximizing rewards, it is often unclear what strategies they use to do so. In this paper, we take a step toward explaining deep RL agents through a case study using Atari 2600 environments. In particular, we focus on using saliency maps to understand how an agent learns and executes a policy. We introduce a method for generating useful saliency maps and use it to show 1) what strong agents attend to, 2) whether agents are making decisions for the right or wrong reasons, and 3) how agents evolve during learning. We also test ou","authors_text":"Alan Fern, Anurag Koul, Jonathan Dodge, Sam Greydanus","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2017-10-31T23:03:17Z","title":"Visualizing and Understanding Atari Agents"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1711.00138","kind":"arxiv","version":5},"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:fa15f6fd696ef1c42d3bf8f906550a3b1f818c5972584c74f7e1537d55bceec9","target":"record","created_at":"2026-05-18T00:06:05Z","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":"161490d60f5586ae63dfe10db392c79a18bcd4292cab00dd9f480a7f7a0ffa0b","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2017-10-31T23:03:17Z","title_canon_sha256":"9ca38b59837320715bbb374620a7e5dc0fa7da050e2c3caba0003867b12351c9"},"schema_version":"1.0","source":{"id":"1711.00138","kind":"arxiv","version":5}},"canonical_sha256":"a525d2f179cc619b4cac48397d8bbbadb266afac7c47a95230df4b622cbfd3ac","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a525d2f179cc619b4cac48397d8bbbadb266afac7c47a95230df4b622cbfd3ac","first_computed_at":"2026-05-18T00:06:05.810269Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T00:06:05.810269Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"RyKNi36iSK38TuiO7jaaZ7jrwljedfA9BkrDs3KpcLW5f8Jc2lYNCHTy34LsLzngCkaMzEAtF8q7VCN9UQcXAg==","signature_status":"signed_v1","signed_at":"2026-05-18T00:06:05.810882Z","signed_message":"canonical_sha256_bytes"},"source_id":"1711.00138","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:fa15f6fd696ef1c42d3bf8f906550a3b1f818c5972584c74f7e1537d55bceec9","sha256:1a996eab4dfaf3f7b0b3fd53d0524ad52ab5b278a48863e5cebbe6f3182ca1a2"],"state_sha256":"4a4f7eb700efae0a187537f93d2209c7244f883700a65205fec6cef22752b40d"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"w/0ik0mW1Br5oooohFrQb8dU9jp2ZeFLtXm4r0V6RtyItvUbvBl8pi5n2PHgrUdmtNRx6NlHP+WIe3HT217vDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T17:45:25.101754Z","bundle_sha256":"e8f083d0933be383f015bc83f43076d287879cca6e0a806769bccd1e17bcff80"}}