{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:2ZKEHNEVC4UEXHFPOQMTTKD5PT","short_pith_number":"pith:2ZKEHNEV","schema_version":"1.0","canonical_sha256":"d65443b49517284b9caf741939a87d7cdf413baec9ed0c155ed4599049783e5e","source":{"kind":"arxiv","id":"2202.04836","version":2},"attestation_state":"computed","paper":{"title":"Deconstructing the Inductive Biases of Hamiltonian Neural Networks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.DS","physics.data-an","stat.ML"],"primary_cat":"cs.LG","authors_text":"Andrew Gordon Wilson, Marc Finzi, Nate Gruver, Samuel Stanton","submitted_at":"2022-02-10T05:05:38Z","abstract_excerpt":"Physics-inspired neural networks (NNs), such as Hamiltonian or Lagrangian NNs, dramatically outperform other learned dynamics models by leveraging strong inductive biases. These models, however, are challenging to apply to many real world systems, such as those that don't conserve energy or contain contacts, a common setting for robotics and reinforcement learning. In this paper, we examine the inductive biases that make physics-inspired models successful in practice. We show that, contrary to conventional wisdom, the improved generalization of HNNs is the result of modeling acceleration direc"},"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":"2202.04836","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-02-10T05:05:38Z","cross_cats_sorted":["math.DS","physics.data-an","stat.ML"],"title_canon_sha256":"588497ca7fb7a290f15332c2dbe323fc108035291c0bb7aa528d81c48309d617","abstract_canon_sha256":"4d24da1609ac54fb959848e58e72fc963e3a2724027fbbb44b650873751995bd"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:56:18.454513Z","signature_b64":"ICt4mY5t8HaZ87bh7Ru7M2ewioUXnFKIinNomkkSdjKGpTuLXOJLpfiFM4Y4WYFhYej7JfSHXwe5NxQ6MWYWAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d65443b49517284b9caf741939a87d7cdf413baec9ed0c155ed4599049783e5e","last_reissued_at":"2026-07-05T03:56:18.453937Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:56:18.453937Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Deconstructing the Inductive Biases of Hamiltonian Neural Networks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.DS","physics.data-an","stat.ML"],"primary_cat":"cs.LG","authors_text":"Andrew Gordon Wilson, Marc Finzi, Nate Gruver, Samuel Stanton","submitted_at":"2022-02-10T05:05:38Z","abstract_excerpt":"Physics-inspired neural networks (NNs), such as Hamiltonian or Lagrangian NNs, dramatically outperform other learned dynamics models by leveraging strong inductive biases. These models, however, are challenging to apply to many real world systems, such as those that don't conserve energy or contain contacts, a common setting for robotics and reinforcement learning. In this paper, we examine the inductive biases that make physics-inspired models successful in practice. We show that, contrary to conventional wisdom, the improved generalization of HNNs is the result of modeling acceleration direc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2202.04836","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/2202.04836/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":"2202.04836","created_at":"2026-07-05T03:56:18.453996+00:00"},{"alias_kind":"arxiv_version","alias_value":"2202.04836v2","created_at":"2026-07-05T03:56:18.453996+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2202.04836","created_at":"2026-07-05T03:56:18.453996+00:00"},{"alias_kind":"pith_short_12","alias_value":"2ZKEHNEVC4UE","created_at":"2026-07-05T03:56:18.453996+00:00"},{"alias_kind":"pith_short_16","alias_value":"2ZKEHNEVC4UEXHFP","created_at":"2026-07-05T03:56:18.453996+00:00"},{"alias_kind":"pith_short_8","alias_value":"2ZKEHNEV","created_at":"2026-07-05T03:56:18.453996+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.20515","citing_title":"Semi-Explicit Neural DAEs: Learning Long-Horizon Dynamical Systems with Algebraic Constraints","ref_index":1986,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2ZKEHNEVC4UEXHFPOQMTTKD5PT","json":"https://pith.science/pith/2ZKEHNEVC4UEXHFPOQMTTKD5PT.json","graph_json":"https://pith.science/api/pith-number/2ZKEHNEVC4UEXHFPOQMTTKD5PT/graph.json","events_json":"https://pith.science/api/pith-number/2ZKEHNEVC4UEXHFPOQMTTKD5PT/events.json","paper":"https://pith.science/paper/2ZKEHNEV"},"agent_actions":{"view_html":"https://pith.science/pith/2ZKEHNEVC4UEXHFPOQMTTKD5PT","download_json":"https://pith.science/pith/2ZKEHNEVC4UEXHFPOQMTTKD5PT.json","view_paper":"https://pith.science/paper/2ZKEHNEV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2202.04836&json=true","fetch_graph":"https://pith.science/api/pith-number/2ZKEHNEVC4UEXHFPOQMTTKD5PT/graph.json","fetch_events":"https://pith.science/api/pith-number/2ZKEHNEVC4UEXHFPOQMTTKD5PT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2ZKEHNEVC4UEXHFPOQMTTKD5PT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2ZKEHNEVC4UEXHFPOQMTTKD5PT/action/storage_attestation","attest_author":"https://pith.science/pith/2ZKEHNEVC4UEXHFPOQMTTKD5PT/action/author_attestation","sign_citation":"https://pith.science/pith/2ZKEHNEVC4UEXHFPOQMTTKD5PT/action/citation_signature","submit_replication":"https://pith.science/pith/2ZKEHNEVC4UEXHFPOQMTTKD5PT/action/replication_record"}},"created_at":"2026-07-05T03:56:18.453996+00:00","updated_at":"2026-07-05T03:56:18.453996+00:00"}