{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:UE73HWMQKB3L4DBO22MA5V7NGF","short_pith_number":"pith:UE73HWMQ","schema_version":"1.0","canonical_sha256":"a13fb3d9905076be0c2ed6980ed7ed3168dc6f2497fe254f31a4ec2305ec0f28","source":{"kind":"arxiv","id":"2003.06016","version":2},"attestation_state":"computed","paper":{"title":"Invariant Causal Prediction for Block MDPs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Amy Zhang, Angelos Filos, Clare Lyle, Doina Precup, Joelle Pineau, Marta Kwiatkowska, Shagun Sodhani, Yarin Gal","submitted_at":"2020-03-12T21:03:01Z","abstract_excerpt":"Generalization across environments is critical to the successful application of reinforcement learning algorithms to real-world challenges. In this paper, we consider the problem of learning abstractions that generalize in block MDPs, families of environments with a shared latent state space and dynamics structure over that latent space, but varying observations. We leverage tools from causal inference to propose a method of invariant prediction to learn model-irrelevance state abstractions (MISA) that generalize to novel observations in the multi-environment setting. We prove that for certain"},"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":"2003.06016","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-03-12T21:03:01Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"c00b02f29b4e8dc614b3a110ef52176d5eb5043fd4dbd526ecf2caeb08b17b09","abstract_canon_sha256":"07285f824d3ad3532cc221d3b472c5773289dae57957eaf6e540c8d39eb0abec"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:09:41.723984Z","signature_b64":"GOt0N3WL+W/BYnHraIxChBTPJtZrg7YDLzibx856oIDkozDDuDz1hdo6AJx9U4nhen8/AmvCtfXD1Ii31V7DDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a13fb3d9905076be0c2ed6980ed7ed3168dc6f2497fe254f31a4ec2305ec0f28","last_reissued_at":"2026-07-05T01:09:41.723493Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:09:41.723493Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Invariant Causal Prediction for Block MDPs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Amy Zhang, Angelos Filos, Clare Lyle, Doina Precup, Joelle Pineau, Marta Kwiatkowska, Shagun Sodhani, Yarin Gal","submitted_at":"2020-03-12T21:03:01Z","abstract_excerpt":"Generalization across environments is critical to the successful application of reinforcement learning algorithms to real-world challenges. In this paper, we consider the problem of learning abstractions that generalize in block MDPs, families of environments with a shared latent state space and dynamics structure over that latent space, but varying observations. We leverage tools from causal inference to propose a method of invariant prediction to learn model-irrelevance state abstractions (MISA) that generalize to novel observations in the multi-environment setting. We prove that for certain"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2003.06016","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/2003.06016/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":"2003.06016","created_at":"2026-07-05T01:09:41.723566+00:00"},{"alias_kind":"arxiv_version","alias_value":"2003.06016v2","created_at":"2026-07-05T01:09:41.723566+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2003.06016","created_at":"2026-07-05T01:09:41.723566+00:00"},{"alias_kind":"pith_short_12","alias_value":"UE73HWMQKB3L","created_at":"2026-07-05T01:09:41.723566+00:00"},{"alias_kind":"pith_short_16","alias_value":"UE73HWMQKB3L4DBO","created_at":"2026-07-05T01:09:41.723566+00:00"},{"alias_kind":"pith_short_8","alias_value":"UE73HWMQ","created_at":"2026-07-05T01:09:41.723566+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.08361","citing_title":"Modeling Unseen Environments with Language-guided Composable Causal Components in Reinforcement Learning","ref_index":2022,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UE73HWMQKB3L4DBO22MA5V7NGF","json":"https://pith.science/pith/UE73HWMQKB3L4DBO22MA5V7NGF.json","graph_json":"https://pith.science/api/pith-number/UE73HWMQKB3L4DBO22MA5V7NGF/graph.json","events_json":"https://pith.science/api/pith-number/UE73HWMQKB3L4DBO22MA5V7NGF/events.json","paper":"https://pith.science/paper/UE73HWMQ"},"agent_actions":{"view_html":"https://pith.science/pith/UE73HWMQKB3L4DBO22MA5V7NGF","download_json":"https://pith.science/pith/UE73HWMQKB3L4DBO22MA5V7NGF.json","view_paper":"https://pith.science/paper/UE73HWMQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2003.06016&json=true","fetch_graph":"https://pith.science/api/pith-number/UE73HWMQKB3L4DBO22MA5V7NGF/graph.json","fetch_events":"https://pith.science/api/pith-number/UE73HWMQKB3L4DBO22MA5V7NGF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UE73HWMQKB3L4DBO22MA5V7NGF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UE73HWMQKB3L4DBO22MA5V7NGF/action/storage_attestation","attest_author":"https://pith.science/pith/UE73HWMQKB3L4DBO22MA5V7NGF/action/author_attestation","sign_citation":"https://pith.science/pith/UE73HWMQKB3L4DBO22MA5V7NGF/action/citation_signature","submit_replication":"https://pith.science/pith/UE73HWMQKB3L4DBO22MA5V7NGF/action/replication_record"}},"created_at":"2026-07-05T01:09:41.723566+00:00","updated_at":"2026-07-05T01:09:41.723566+00:00"}