{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:BHNLPCNA7RONF4I56HBKGHVN25","short_pith_number":"pith:BHNLPCNA","schema_version":"1.0","canonical_sha256":"09dab789a0fc5cd2f11df1c2a31eadd76299812f8957970e2fa240a4fcfe08a3","source":{"kind":"arxiv","id":"2203.09637","version":1},"attestation_state":"computed","paper":{"title":"Investigating Compounding Prediction Errors in Learned Dynamics Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Kristofer Pister, Nathan Lambert, Roberto Calandra","submitted_at":"2022-03-17T22:24:38Z","abstract_excerpt":"Accurately predicting the consequences of agents' actions is a key prerequisite for planning in robotic control. Model-based reinforcement learning (MBRL) is one paradigm which relies on the iterative learning and prediction of state-action transitions to solve a task. Deep MBRL has become a popular candidate, using a neural network to learn a dynamics model that predicts with each pass from high-dimensional states to actions. These \"one-step\" predictions are known to become inaccurate over longer horizons of composed prediction - called the compounding error problem. Given the prevalence of t"},"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":"2203.09637","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-03-17T22:24:38Z","cross_cats_sorted":[],"title_canon_sha256":"e002fabf069603511c32c7c498558725df5b2351b731ad377ad5469038613130","abstract_canon_sha256":"23bd9ea734820bf63613493d0e95a0a1dbfcd88701872dbb49926e066d29c19f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:06:22.019536Z","signature_b64":"K3aEqpt7YFwFvmtUt9WN9Mc06Ww/cbq8S+h1ZzrOpoh3nEA8cDa12QbJpKXliThJdKVLb+S7zJMLOu9B7YlsAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"09dab789a0fc5cd2f11df1c2a31eadd76299812f8957970e2fa240a4fcfe08a3","last_reissued_at":"2026-07-05T04:06:22.019089Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:06:22.019089Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Investigating Compounding Prediction Errors in Learned Dynamics Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Kristofer Pister, Nathan Lambert, Roberto Calandra","submitted_at":"2022-03-17T22:24:38Z","abstract_excerpt":"Accurately predicting the consequences of agents' actions is a key prerequisite for planning in robotic control. Model-based reinforcement learning (MBRL) is one paradigm which relies on the iterative learning and prediction of state-action transitions to solve a task. Deep MBRL has become a popular candidate, using a neural network to learn a dynamics model that predicts with each pass from high-dimensional states to actions. These \"one-step\" predictions are known to become inaccurate over longer horizons of composed prediction - called the compounding error problem. Given the prevalence of t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.09637","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/2203.09637/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":"2203.09637","created_at":"2026-07-05T04:06:22.019147+00:00"},{"alias_kind":"arxiv_version","alias_value":"2203.09637v1","created_at":"2026-07-05T04:06:22.019147+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.09637","created_at":"2026-07-05T04:06:22.019147+00:00"},{"alias_kind":"pith_short_12","alias_value":"BHNLPCNA7RON","created_at":"2026-07-05T04:06:22.019147+00:00"},{"alias_kind":"pith_short_16","alias_value":"BHNLPCNA7RONF4I5","created_at":"2026-07-05T04:06:22.019147+00:00"},{"alias_kind":"pith_short_8","alias_value":"BHNLPCNA","created_at":"2026-07-05T04:06:22.019147+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.06401","citing_title":"A Definition and Roadmap for World Models","ref_index":48,"is_internal_anchor":true},{"citing_arxiv_id":"2606.09640","citing_title":"Physics-Aware Sparse Learning and Selective Online Adaptation for Euler-Lagrange Robot Dynamics","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14805","citing_title":"Learning Cross-Coupled and Regime Dependent Dynamics for Aerial Manipulation","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26282","citing_title":"Scaling World-Model Reinforcement Learning Through Diffusion Policy Optimization","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2506.08902","citing_title":"Intention-Conditioned Flow Occupancy Models","ref_index":60,"is_internal_anchor":false},{"citing_arxiv_id":"2512.04341","citing_title":"Long-Horizon Model-Based Offline Reinforcement Learning Without Explicit Conservatism","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BHNLPCNA7RONF4I56HBKGHVN25","json":"https://pith.science/pith/BHNLPCNA7RONF4I56HBKGHVN25.json","graph_json":"https://pith.science/api/pith-number/BHNLPCNA7RONF4I56HBKGHVN25/graph.json","events_json":"https://pith.science/api/pith-number/BHNLPCNA7RONF4I56HBKGHVN25/events.json","paper":"https://pith.science/paper/BHNLPCNA"},"agent_actions":{"view_html":"https://pith.science/pith/BHNLPCNA7RONF4I56HBKGHVN25","download_json":"https://pith.science/pith/BHNLPCNA7RONF4I56HBKGHVN25.json","view_paper":"https://pith.science/paper/BHNLPCNA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2203.09637&json=true","fetch_graph":"https://pith.science/api/pith-number/BHNLPCNA7RONF4I56HBKGHVN25/graph.json","fetch_events":"https://pith.science/api/pith-number/BHNLPCNA7RONF4I56HBKGHVN25/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BHNLPCNA7RONF4I56HBKGHVN25/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BHNLPCNA7RONF4I56HBKGHVN25/action/storage_attestation","attest_author":"https://pith.science/pith/BHNLPCNA7RONF4I56HBKGHVN25/action/author_attestation","sign_citation":"https://pith.science/pith/BHNLPCNA7RONF4I56HBKGHVN25/action/citation_signature","submit_replication":"https://pith.science/pith/BHNLPCNA7RONF4I56HBKGHVN25/action/replication_record"}},"created_at":"2026-07-05T04:06:22.019147+00:00","updated_at":"2026-07-05T04:06:22.019147+00:00"}