{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NPUXPVLGJZYSRJJAKK2SWP3TJH","short_pith_number":"pith:NPUXPVLG","schema_version":"1.0","canonical_sha256":"6be977d5664e7128a52052b52b3f7349cbddaf8e0baca1d46486f6d8eaff2ed7","source":{"kind":"arxiv","id":"2602.19373","version":3},"attestation_state":"computed","paper":{"title":"Stable Deep Reinforcement Learning via Isotropic Gaussian Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Aaron Courville, Ali Saheb Pasand, Johan Obando-Ceron, Pablo Samuel Castro, Pouya Bashivan","submitted_at":"2026-02-22T22:55:27Z","abstract_excerpt":"Deep reinforcement learning systems often suffer from unstable training dynamics due to non-stationarity, where learning objectives and data distributions evolve over time. We show that under non-stationary targets, isotropic Gaussian embeddings are provably advantageous. In particular, they induce stable tracking of time-varying targets for linear readouts, achieve maximal entropy under a fixed variance budget, and encourage a balanced use of all representational dimensions--all of which enable agents to be more adaptive and stable. Building on this insight, we propose the use of Sketched Iso"},"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":"2602.19373","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-02-22T22:55:27Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9a17f363bb327ac1c13c9300413b6f7f1c9e59a815dd4c40fbffbfa94d1d9d2e","abstract_canon_sha256":"6c087ae2f40aee568e6a057af6ed037ded08987a486c4f8c2e103067846a702c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T01:14:35.724726Z","signature_b64":"ICc6e8L0T9ayye2sZkEgJBpcm4o7zO5w4i8MGKCmPY0OOsBjEKXTvD581R2uui+BVx4hiMnBtTmFwVpY/VZ5DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6be977d5664e7128a52052b52b3f7349cbddaf8e0baca1d46486f6d8eaff2ed7","last_reissued_at":"2026-06-05T01:14:35.724052Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T01:14:35.724052Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Stable Deep Reinforcement Learning via Isotropic Gaussian Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Aaron Courville, Ali Saheb Pasand, Johan Obando-Ceron, Pablo Samuel Castro, Pouya Bashivan","submitted_at":"2026-02-22T22:55:27Z","abstract_excerpt":"Deep reinforcement learning systems often suffer from unstable training dynamics due to non-stationarity, where learning objectives and data distributions evolve over time. We show that under non-stationary targets, isotropic Gaussian embeddings are provably advantageous. In particular, they induce stable tracking of time-varying targets for linear readouts, achieve maximal entropy under a fixed variance budget, and encourage a balanced use of all representational dimensions--all of which enable agents to be more adaptive and stable. Building on this insight, we propose the use of Sketched Iso"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.19373","kind":"arxiv","version":3},"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/2602.19373/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":"2602.19373","created_at":"2026-06-05T01:14:35.724119+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.19373v3","created_at":"2026-06-05T01:14:35.724119+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.19373","created_at":"2026-06-05T01:14:35.724119+00:00"},{"alias_kind":"pith_short_12","alias_value":"NPUXPVLGJZYS","created_at":"2026-06-05T01:14:35.724119+00:00"},{"alias_kind":"pith_short_16","alias_value":"NPUXPVLGJZYSRJJA","created_at":"2026-06-05T01:14:35.724119+00:00"},{"alias_kind":"pith_short_8","alias_value":"NPUXPVLG","created_at":"2026-06-05T01:14:35.724119+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/NPUXPVLGJZYSRJJAKK2SWP3TJH","json":"https://pith.science/pith/NPUXPVLGJZYSRJJAKK2SWP3TJH.json","graph_json":"https://pith.science/api/pith-number/NPUXPVLGJZYSRJJAKK2SWP3TJH/graph.json","events_json":"https://pith.science/api/pith-number/NPUXPVLGJZYSRJJAKK2SWP3TJH/events.json","paper":"https://pith.science/paper/NPUXPVLG"},"agent_actions":{"view_html":"https://pith.science/pith/NPUXPVLGJZYSRJJAKK2SWP3TJH","download_json":"https://pith.science/pith/NPUXPVLGJZYSRJJAKK2SWP3TJH.json","view_paper":"https://pith.science/paper/NPUXPVLG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.19373&json=true","fetch_graph":"https://pith.science/api/pith-number/NPUXPVLGJZYSRJJAKK2SWP3TJH/graph.json","fetch_events":"https://pith.science/api/pith-number/NPUXPVLGJZYSRJJAKK2SWP3TJH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NPUXPVLGJZYSRJJAKK2SWP3TJH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NPUXPVLGJZYSRJJAKK2SWP3TJH/action/storage_attestation","attest_author":"https://pith.science/pith/NPUXPVLGJZYSRJJAKK2SWP3TJH/action/author_attestation","sign_citation":"https://pith.science/pith/NPUXPVLGJZYSRJJAKK2SWP3TJH/action/citation_signature","submit_replication":"https://pith.science/pith/NPUXPVLGJZYSRJJAKK2SWP3TJH/action/replication_record"}},"created_at":"2026-06-05T01:14:35.724119+00:00","updated_at":"2026-06-05T01:14:35.724119+00:00"}