{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:VHTLMS364QWJGYMGJSHLEBAF73","short_pith_number":"pith:VHTLMS36","schema_version":"1.0","canonical_sha256":"a9e6b64b7ee42c9361864c8eb20405feee90b164f044c6e6754e18ce0728f7ea","source":{"kind":"arxiv","id":"2305.14133","version":2},"attestation_state":"computed","paper":{"title":"Conditional Mutual Information for Disentangled Representations in Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Josiah P. Hanna, Kevin Sebastian Luck, Mhairi Dunion, Stefano V. Albrecht, Trevor McInroe","submitted_at":"2023-05-23T14:56:19Z","abstract_excerpt":"Reinforcement Learning (RL) environments can produce training data with spurious correlations between features due to the amount of training data or its limited feature coverage. This can lead to RL agents encoding these misleading correlations in their latent representation, preventing the agent from generalising if the correlation changes within the environment or when deployed in the real world. Disentangled representations can improve robustness, but existing disentanglement techniques that minimise mutual information between features require independent features, thus they cannot disentan"},"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":"2305.14133","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-05-23T14:56:19Z","cross_cats_sorted":[],"title_canon_sha256":"6015e1a2b8345ec274804ee754b8551c311da02b709fccb8e859ad523f2fddfe","abstract_canon_sha256":"ac7bb575ccccd99b5dffab1e95461b6c3f9194a9e0540d5ffc0061d0e8963b4c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:00:08.176661Z","signature_b64":"SA51rrWk9Fqkw1wnKMovlrAhaY2Db9afiph9gCDRCmcAuQ3Hzx3vB7XpX7zWsUqO7p5fZJgg9Td6chZYhYw5Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a9e6b64b7ee42c9361864c8eb20405feee90b164f044c6e6754e18ce0728f7ea","last_reissued_at":"2026-07-05T07:00:08.176192Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:00:08.176192Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Conditional Mutual Information for Disentangled Representations in Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Josiah P. Hanna, Kevin Sebastian Luck, Mhairi Dunion, Stefano V. Albrecht, Trevor McInroe","submitted_at":"2023-05-23T14:56:19Z","abstract_excerpt":"Reinforcement Learning (RL) environments can produce training data with spurious correlations between features due to the amount of training data or its limited feature coverage. This can lead to RL agents encoding these misleading correlations in their latent representation, preventing the agent from generalising if the correlation changes within the environment or when deployed in the real world. Disentangled representations can improve robustness, but existing disentanglement techniques that minimise mutual information between features require independent features, thus they cannot disentan"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.14133","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/2305.14133/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":"2305.14133","created_at":"2026-07-05T07:00:08.176249+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.14133v2","created_at":"2026-07-05T07:00:08.176249+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.14133","created_at":"2026-07-05T07:00:08.176249+00:00"},{"alias_kind":"pith_short_12","alias_value":"VHTLMS364QWJ","created_at":"2026-07-05T07:00:08.176249+00:00"},{"alias_kind":"pith_short_16","alias_value":"VHTLMS364QWJGYMG","created_at":"2026-07-05T07:00:08.176249+00:00"},{"alias_kind":"pith_short_8","alias_value":"VHTLMS36","created_at":"2026-07-05T07:00:08.176249+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/VHTLMS364QWJGYMGJSHLEBAF73","json":"https://pith.science/pith/VHTLMS364QWJGYMGJSHLEBAF73.json","graph_json":"https://pith.science/api/pith-number/VHTLMS364QWJGYMGJSHLEBAF73/graph.json","events_json":"https://pith.science/api/pith-number/VHTLMS364QWJGYMGJSHLEBAF73/events.json","paper":"https://pith.science/paper/VHTLMS36"},"agent_actions":{"view_html":"https://pith.science/pith/VHTLMS364QWJGYMGJSHLEBAF73","download_json":"https://pith.science/pith/VHTLMS364QWJGYMGJSHLEBAF73.json","view_paper":"https://pith.science/paper/VHTLMS36","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.14133&json=true","fetch_graph":"https://pith.science/api/pith-number/VHTLMS364QWJGYMGJSHLEBAF73/graph.json","fetch_events":"https://pith.science/api/pith-number/VHTLMS364QWJGYMGJSHLEBAF73/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VHTLMS364QWJGYMGJSHLEBAF73/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VHTLMS364QWJGYMGJSHLEBAF73/action/storage_attestation","attest_author":"https://pith.science/pith/VHTLMS364QWJGYMGJSHLEBAF73/action/author_attestation","sign_citation":"https://pith.science/pith/VHTLMS364QWJGYMGJSHLEBAF73/action/citation_signature","submit_replication":"https://pith.science/pith/VHTLMS364QWJGYMGJSHLEBAF73/action/replication_record"}},"created_at":"2026-07-05T07:00:08.176249+00:00","updated_at":"2026-07-05T07:00:08.176249+00:00"}