{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:C5OAVKATGIQL7Z2AGR3ZFKH3DJ","short_pith_number":"pith:C5OAVKAT","schema_version":"1.0","canonical_sha256":"175c0aa8133220bfe740347792a8fb1a7f9a19d0866d97850b939c573dfcf494","source":{"kind":"arxiv","id":"2605.12823","version":1},"attestation_state":"computed","paper":{"title":"Hessian Matching for Machine-Learned Coarse-Grained Molecular Dynamics","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Adding stochastic Hessian-vector product matching to force matching improves coarse-grained molecular dynamics models by incorporating curvature information.","cross_cats":["physics.chem-ph","physics.comp-ph","q-bio.BM"],"primary_cat":"cs.LG","authors_text":"Ashwin Lokapally, Kevin Bachelor, Razvan Marinescu, Sanjit Shashi, Sanya Murdeshwar, William Noid","submitted_at":"2026-05-12T23:46:38Z","abstract_excerpt":"Coarse-grained (CG) molecular dynamics enables simulations of atomic systems such as biomolecules at timescales inaccessible to all-atom (AA) methods, but existing CG neural potentials trained via force matching capture only the gradient of the free-energy surface, leaving its curvature unconstrained. We introduce a framework that augments force matching with stochastic Hessian-vector product (HVP) matching, instilling second-order curvature information into CG potentials without constructing the full Hessian. We derive a decomposition of the target CG Hessian into a model-independent projecte"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.12823","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-12T23:46:38Z","cross_cats_sorted":["physics.chem-ph","physics.comp-ph","q-bio.BM"],"title_canon_sha256":"32da991e95ad9a3b3262ff36fdaf295cae24357fb9514c1095b649f772e00907","abstract_canon_sha256":"eb975d273f1af5eff956caa1781ce3ce78256fbc89e4e3890d6c7834b14d145d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:09:12.195540Z","signature_b64":"6/5b55zXKTQqDA6tHzahAS3Tj+MevVSmP3DFlERw2UPRKbHtXGiqXU4vzH36A/7xy2fasOGWx+FlupHkW8Y+AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"175c0aa8133220bfe740347792a8fb1a7f9a19d0866d97850b939c573dfcf494","last_reissued_at":"2026-05-18T03:09:12.194820Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:09:12.194820Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hessian Matching for Machine-Learned Coarse-Grained Molecular Dynamics","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Adding stochastic Hessian-vector product matching to force matching improves coarse-grained molecular dynamics models by incorporating curvature information.","cross_cats":["physics.chem-ph","physics.comp-ph","q-bio.BM"],"primary_cat":"cs.LG","authors_text":"Ashwin Lokapally, Kevin Bachelor, Razvan Marinescu, Sanjit Shashi, Sanya Murdeshwar, William Noid","submitted_at":"2026-05-12T23:46:38Z","abstract_excerpt":"Coarse-grained (CG) molecular dynamics enables simulations of atomic systems such as biomolecules at timescales inaccessible to all-atom (AA) methods, but existing CG neural potentials trained via force matching capture only the gradient of the free-energy surface, leaving its curvature unconstrained. We introduce a framework that augments force matching with stochastic Hessian-vector product (HVP) matching, instilling second-order curvature information into CG potentials without constructing the full Hessian. We derive a decomposition of the target CG Hessian into a model-independent projecte"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"HVP matching outperforms plain force matching on 8 of 9 proteins on slow-mode metrics, with reductions of up to 85% in the Kullback--Leibler divergence between the CG and reference distributions along the slowest collective mode of the largest protein.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The decomposition of the target CG Hessian into a model-independent projected AA Hessian plus a model-dependent covariance correction remains valid and the stochastic HVP estimator stays unbiased throughout training.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Stochastic Hessian-vector product matching augments force matching to instill second-order curvature into coarse-grained molecular dynamics potentials, outperforming force matching on slow-mode metrics for 8 of 9 test proteins.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Adding stochastic Hessian-vector product matching to force matching improves coarse-grained molecular dynamics models by incorporating curvature information.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"515881567f4c384b6643fd2cf7030131e20416335d6749c0cb90935b74b4674c"},"source":{"id":"2605.12823","kind":"arxiv","version":1},"verdict":{"id":"ced17674-041f-44d6-a879-2a385ddd4bc8","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:50:39.104470Z","strongest_claim":"HVP matching outperforms plain force matching on 8 of 9 proteins on slow-mode metrics, with reductions of up to 85% in the Kullback--Leibler divergence between the CG and reference distributions along the slowest collective mode of the largest protein.","one_line_summary":"Stochastic Hessian-vector product matching augments force matching to instill second-order curvature into coarse-grained molecular dynamics potentials, outperforming force matching on slow-mode metrics for 8 of 9 test proteins.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The decomposition of the target CG Hessian into a model-independent projected AA Hessian plus a model-dependent covariance correction remains valid and the stochastic HVP estimator stays unbiased throughout training.","pith_extraction_headline":"Adding stochastic Hessian-vector product matching to force matching improves coarse-grained molecular dynamics models by incorporating curvature information."},"references":{"count":32,"sample":[{"doi":"","year":2009,"title":"David E. Shaw et al. Millisecond-scale molecular dynamics simulations on anton. InProceed- ings of the Conference on High Performance Computing Networking, Storage and Analysis, SC ’09, page 1–11. ACM","work_id":"1f5b5ee6-df99-4188-96ad-c4a62f5cd374","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"David E. Shaw et al. Anton 3: Twenty microseconds of molecular dynamics simulation before lunch. InSC21: International Conference for High Performance Computing, Networking, Storage and Analysis, page","work_id":"e50819a6-da72-4a92-b7de-8c2265e934ab","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2013,"title":"W. G. Noid. Perspective: Coarse-grained models for biomolecular systems.The Journal of Chemical Physics, 139(9), September 2013","work_id":"a09b386a-2953-4759-a50f-e29ce09b0b28","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2016,"title":"Coarse-grained protein models and their applications.Chemical Reviews, 116(14):7898–7936, June 2016","work_id":"1245d719-86bb-4ec9-af13-2b74b03f1dab","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1994,"title":"F. Ercolessi and J. B. Adams. Interatomic potentials from first-principles calculations: The force-matching method.Europhysics Letters, 26(8):583–588, 1994","work_id":"99f1ecc3-0a85-427a-84ee-e008ac6df957","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":32,"snapshot_sha256":"ae91b902a5dee3666a059c3381ca7bf352c7f776ce354c511a41a6742fcaa29e","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"97b3ea85d005e3a0aa26caef444d1559eaf94608c352a1b0ca1566a02bc2a242"},"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":"2605.12823","created_at":"2026-05-18T03:09:12.194962+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.12823v1","created_at":"2026-05-18T03:09:12.194962+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.12823","created_at":"2026-05-18T03:09:12.194962+00:00"},{"alias_kind":"pith_short_12","alias_value":"C5OAVKATGIQL","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"C5OAVKATGIQL7Z2A","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"C5OAVKAT","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/C5OAVKATGIQL7Z2AGR3ZFKH3DJ","json":"https://pith.science/pith/C5OAVKATGIQL7Z2AGR3ZFKH3DJ.json","graph_json":"https://pith.science/api/pith-number/C5OAVKATGIQL7Z2AGR3ZFKH3DJ/graph.json","events_json":"https://pith.science/api/pith-number/C5OAVKATGIQL7Z2AGR3ZFKH3DJ/events.json","paper":"https://pith.science/paper/C5OAVKAT"},"agent_actions":{"view_html":"https://pith.science/pith/C5OAVKATGIQL7Z2AGR3ZFKH3DJ","download_json":"https://pith.science/pith/C5OAVKATGIQL7Z2AGR3ZFKH3DJ.json","view_paper":"https://pith.science/paper/C5OAVKAT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.12823&json=true","fetch_graph":"https://pith.science/api/pith-number/C5OAVKATGIQL7Z2AGR3ZFKH3DJ/graph.json","fetch_events":"https://pith.science/api/pith-number/C5OAVKATGIQL7Z2AGR3ZFKH3DJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/C5OAVKATGIQL7Z2AGR3ZFKH3DJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/C5OAVKATGIQL7Z2AGR3ZFKH3DJ/action/storage_attestation","attest_author":"https://pith.science/pith/C5OAVKATGIQL7Z2AGR3ZFKH3DJ/action/author_attestation","sign_citation":"https://pith.science/pith/C5OAVKATGIQL7Z2AGR3ZFKH3DJ/action/citation_signature","submit_replication":"https://pith.science/pith/C5OAVKATGIQL7Z2AGR3ZFKH3DJ/action/replication_record"}},"created_at":"2026-05-18T03:09:12.194962+00:00","updated_at":"2026-05-18T03:09:12.194962+00:00"}