{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:VLADR4IFIG6RE66QDNSS2PLLLK","short_pith_number":"pith:VLADR4IF","schema_version":"1.0","canonical_sha256":"aac038f10541bd127bd01b652d3d6b5a80097b00bc6c76b36c539a9b9640a0b4","source":{"kind":"arxiv","id":"2507.19437","version":1},"attestation_state":"computed","paper":{"title":"Observations Meet Actions: Learning Control-Sufficient Representations for Robust Policy Generalization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hongpeng Cao, Marco Caccamo, Naira Hovakimyan, Yuliang Gu","submitted_at":"2025-07-25T17:08:16Z","abstract_excerpt":"Capturing latent variations (\"contexts\") is key to deploying reinforcement-learning (RL) agents beyond their training regime. We recast context-based RL as a dual inference-control problem and formally characterize two properties and their hierarchy: observation sufficiency (preserving all predictive information) and control sufficiency (retaining decision-making relevant information). Exploiting this dichotomy, we derive a contextual evidence lower bound(ELBO)-style objective that cleanly separates representation learning from policy learning and optimizes it with Bottlenecked Contextual Poli"},"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":"2507.19437","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-25T17:08:16Z","cross_cats_sorted":[],"title_canon_sha256":"3acb9f59198acae7393e4ccffec300f066362ba7deeab4da00b7bf90c4229382","abstract_canon_sha256":"5b7d0436cab8a2f85af224ee8166f10023a237cff8186e10965184dc98a7ab28"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:43:25.313687Z","signature_b64":"8E1Th1DHlRjZuK/ofpULB33mjYUrupV2BGv57BNfgisC5ZWYNRE0E04H9DHgqtc9E2cURfZMxGtJM0xCKgboBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aac038f10541bd127bd01b652d3d6b5a80097b00bc6c76b36c539a9b9640a0b4","last_reissued_at":"2026-07-05T11:43:25.313165Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:43:25.313165Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Observations Meet Actions: Learning Control-Sufficient Representations for Robust Policy Generalization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hongpeng Cao, Marco Caccamo, Naira Hovakimyan, Yuliang Gu","submitted_at":"2025-07-25T17:08:16Z","abstract_excerpt":"Capturing latent variations (\"contexts\") is key to deploying reinforcement-learning (RL) agents beyond their training regime. We recast context-based RL as a dual inference-control problem and formally characterize two properties and their hierarchy: observation sufficiency (preserving all predictive information) and control sufficiency (retaining decision-making relevant information). Exploiting this dichotomy, we derive a contextual evidence lower bound(ELBO)-style objective that cleanly separates representation learning from policy learning and optimizes it with Bottlenecked Contextual Poli"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.19437","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2507.19437/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":"2507.19437","created_at":"2026-07-05T11:43:25.313226+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.19437v1","created_at":"2026-07-05T11:43:25.313226+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.19437","created_at":"2026-07-05T11:43:25.313226+00:00"},{"alias_kind":"pith_short_12","alias_value":"VLADR4IFIG6R","created_at":"2026-07-05T11:43:25.313226+00:00"},{"alias_kind":"pith_short_16","alias_value":"VLADR4IFIG6RE66Q","created_at":"2026-07-05T11:43:25.313226+00:00"},{"alias_kind":"pith_short_8","alias_value":"VLADR4IF","created_at":"2026-07-05T11:43:25.313226+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/VLADR4IFIG6RE66QDNSS2PLLLK","json":"https://pith.science/pith/VLADR4IFIG6RE66QDNSS2PLLLK.json","graph_json":"https://pith.science/api/pith-number/VLADR4IFIG6RE66QDNSS2PLLLK/graph.json","events_json":"https://pith.science/api/pith-number/VLADR4IFIG6RE66QDNSS2PLLLK/events.json","paper":"https://pith.science/paper/VLADR4IF"},"agent_actions":{"view_html":"https://pith.science/pith/VLADR4IFIG6RE66QDNSS2PLLLK","download_json":"https://pith.science/pith/VLADR4IFIG6RE66QDNSS2PLLLK.json","view_paper":"https://pith.science/paper/VLADR4IF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.19437&json=true","fetch_graph":"https://pith.science/api/pith-number/VLADR4IFIG6RE66QDNSS2PLLLK/graph.json","fetch_events":"https://pith.science/api/pith-number/VLADR4IFIG6RE66QDNSS2PLLLK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VLADR4IFIG6RE66QDNSS2PLLLK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VLADR4IFIG6RE66QDNSS2PLLLK/action/storage_attestation","attest_author":"https://pith.science/pith/VLADR4IFIG6RE66QDNSS2PLLLK/action/author_attestation","sign_citation":"https://pith.science/pith/VLADR4IFIG6RE66QDNSS2PLLLK/action/citation_signature","submit_replication":"https://pith.science/pith/VLADR4IFIG6RE66QDNSS2PLLLK/action/replication_record"}},"created_at":"2026-07-05T11:43:25.313226+00:00","updated_at":"2026-07-05T11:43:25.313226+00:00"}