{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:5WY6XV6B47DCJAQURO56RB3AN5","short_pith_number":"pith:5WY6XV6B","schema_version":"1.0","canonical_sha256":"edb1ebd7c1e7c62482148bbbe887606f6ec1d27a48555d6df4edcb292fad0736","source":{"kind":"arxiv","id":"2402.12971","version":2},"attestation_state":"computed","paper":{"title":"Differentiability in Unrolled Training of Neural Physics Simulators on Transient Dynamics","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"physics.comp-ph","authors_text":"Bjoern List, Kartik Bali, Li-Wei Chen, Nils Thuerey","submitted_at":"2024-02-20T12:40:31Z","abstract_excerpt":"Unrolling training trajectories over time strongly influences the inference accuracy of neural network-augmented physics simulators. We analyze this in three variants of training neural time-steppers. In addition to one-step setups and fully differentiable unrolling, we include a third, less widely used variant: unrolling without temporal gradients. Comparing networks trained with these three modalities disentangles the two dominant effects of unrolling, training distribution shift and long-term gradients. We present detailed study across physical systems, network sizes and architectures, trai"},"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":"2402.12971","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"physics.comp-ph","submitted_at":"2024-02-20T12:40:31Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"2e20208044ab5ef2b15b2b42dc679def19235540558d19ab4abac6c6ceee6160","abstract_canon_sha256":"dad459843afad84ec56ce69f23cee4707cb8ecedb4b5fe16c30ef5e555dd6c6f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:18:23.088447Z","signature_b64":"dqfoQUlVmg+TCl/w2aJAK6EYlGtJpeYfpW5uSIeUHWChSZW8+83u8NjwHYgZBlD5W38AJigQtDyrAfMLeBz8Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"edb1ebd7c1e7c62482148bbbe887606f6ec1d27a48555d6df4edcb292fad0736","last_reissued_at":"2026-07-05T09:18:23.087995Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:18:23.087995Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Differentiability in Unrolled Training of Neural Physics Simulators on Transient Dynamics","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"physics.comp-ph","authors_text":"Bjoern List, Kartik Bali, Li-Wei Chen, Nils Thuerey","submitted_at":"2024-02-20T12:40:31Z","abstract_excerpt":"Unrolling training trajectories over time strongly influences the inference accuracy of neural network-augmented physics simulators. We analyze this in three variants of training neural time-steppers. In addition to one-step setups and fully differentiable unrolling, we include a third, less widely used variant: unrolling without temporal gradients. Comparing networks trained with these three modalities disentangles the two dominant effects of unrolling, training distribution shift and long-term gradients. We present detailed study across physical systems, network sizes and architectures, trai"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.12971","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/2402.12971/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":"2402.12971","created_at":"2026-07-05T09:18:23.088046+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.12971v2","created_at":"2026-07-05T09:18:23.088046+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.12971","created_at":"2026-07-05T09:18:23.088046+00:00"},{"alias_kind":"pith_short_12","alias_value":"5WY6XV6B47DC","created_at":"2026-07-05T09:18:23.088046+00:00"},{"alias_kind":"pith_short_16","alias_value":"5WY6XV6B47DCJAQU","created_at":"2026-07-05T09:18:23.088046+00:00"},{"alias_kind":"pith_short_8","alias_value":"5WY6XV6B","created_at":"2026-07-05T09:18:23.088046+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.27895","citing_title":"Mosaic: A Benchmark Suite for Differentiable Physics Solvers","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2602.05416","citing_title":"Reduced-Order Surrogates for Forced Flexible Mesh Coastal-Ocean Models","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03086","citing_title":"iGENE: A Differentiable Flux-Tube Gyrokinetic Code in TensorFlow","ref_index":71,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5WY6XV6B47DCJAQURO56RB3AN5","json":"https://pith.science/pith/5WY6XV6B47DCJAQURO56RB3AN5.json","graph_json":"https://pith.science/api/pith-number/5WY6XV6B47DCJAQURO56RB3AN5/graph.json","events_json":"https://pith.science/api/pith-number/5WY6XV6B47DCJAQURO56RB3AN5/events.json","paper":"https://pith.science/paper/5WY6XV6B"},"agent_actions":{"view_html":"https://pith.science/pith/5WY6XV6B47DCJAQURO56RB3AN5","download_json":"https://pith.science/pith/5WY6XV6B47DCJAQURO56RB3AN5.json","view_paper":"https://pith.science/paper/5WY6XV6B","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.12971&json=true","fetch_graph":"https://pith.science/api/pith-number/5WY6XV6B47DCJAQURO56RB3AN5/graph.json","fetch_events":"https://pith.science/api/pith-number/5WY6XV6B47DCJAQURO56RB3AN5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5WY6XV6B47DCJAQURO56RB3AN5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5WY6XV6B47DCJAQURO56RB3AN5/action/storage_attestation","attest_author":"https://pith.science/pith/5WY6XV6B47DCJAQURO56RB3AN5/action/author_attestation","sign_citation":"https://pith.science/pith/5WY6XV6B47DCJAQURO56RB3AN5/action/citation_signature","submit_replication":"https://pith.science/pith/5WY6XV6B47DCJAQURO56RB3AN5/action/replication_record"}},"created_at":"2026-07-05T09:18:23.088046+00:00","updated_at":"2026-07-05T09:18:23.088046+00:00"}