{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CYKS52THMHZX6WGLQWJ656GUEO","short_pith_number":"pith:CYKS52TH","schema_version":"1.0","canonical_sha256":"16152eea6761f37f58cb8593eef8d423a2a5c66baacc6cae397b06894b735fdc","source":{"kind":"arxiv","id":"2507.18807","version":1},"attestation_state":"computed","paper":{"title":"Fishers for Free? Approximating the Fisher Information Matrix by Recycling the Squared Gradient Accumulator","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Colin Raffel, Derek Tam, Felix Dangel, Yuxin Li","submitted_at":"2025-07-24T21:10:37Z","abstract_excerpt":"The diagonal of a model's Fisher Information Matrix (the \"Fisher diagonal\") has frequently been used as a way to measure parameter sensitivity. Typically, the Fisher diagonal is estimated via squared sampled gradients of the model's likelihood with respect to its parameters, averaged over a few hundred or thousand examples -- a process which incurs nontrivial computational costs. At the same time, adaptive gradient methods like the ubiquitous Adam optimizer compute a moving average of the squared gradient over the course of training. This paper therefore explores whether an approximation of th"},"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":"2507.18807","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-07-24T21:10:37Z","cross_cats_sorted":[],"title_canon_sha256":"3e6b703e394eaeb3ecc8d3e01306d39a2dbf651c7f2d523928f692ec9a3d8764","abstract_canon_sha256":"a85a89a6f7b616909bbb29010cfcacf8b69aa02e68d0d1e03f0ede288998a1da"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:42:58.630077Z","signature_b64":"17DocsH/bIjbbFf3tlSdIyoPMj42bzh7kCCaMmxVrAfzxw1ktW6r3Z6F5HuqTf3lW+ocFyJ2UpKRcx4+5g81Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"16152eea6761f37f58cb8593eef8d423a2a5c66baacc6cae397b06894b735fdc","last_reissued_at":"2026-07-05T11:42:58.629555Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:42:58.629555Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Fishers for Free? Approximating the Fisher Information Matrix by Recycling the Squared Gradient Accumulator","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Colin Raffel, Derek Tam, Felix Dangel, Yuxin Li","submitted_at":"2025-07-24T21:10:37Z","abstract_excerpt":"The diagonal of a model's Fisher Information Matrix (the \"Fisher diagonal\") has frequently been used as a way to measure parameter sensitivity. Typically, the Fisher diagonal is estimated via squared sampled gradients of the model's likelihood with respect to its parameters, averaged over a few hundred or thousand examples -- a process which incurs nontrivial computational costs. At the same time, adaptive gradient methods like the ubiquitous Adam optimizer compute a moving average of the squared gradient over the course of training. This paper therefore explores whether an approximation of th"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.18807","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/2507.18807/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":"2507.18807","created_at":"2026-07-05T11:42:58.629617+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.18807v1","created_at":"2026-07-05T11:42:58.629617+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.18807","created_at":"2026-07-05T11:42:58.629617+00:00"},{"alias_kind":"pith_short_12","alias_value":"CYKS52THMHZX","created_at":"2026-07-05T11:42:58.629617+00:00"},{"alias_kind":"pith_short_16","alias_value":"CYKS52THMHZX6WGL","created_at":"2026-07-05T11:42:58.629617+00:00"},{"alias_kind":"pith_short_8","alias_value":"CYKS52TH","created_at":"2026-07-05T11:42:58.629617+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03904","citing_title":"MAdam: Metric-Aware Multi-Objective Adam","ref_index":24,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CYKS52THMHZX6WGLQWJ656GUEO","json":"https://pith.science/pith/CYKS52THMHZX6WGLQWJ656GUEO.json","graph_json":"https://pith.science/api/pith-number/CYKS52THMHZX6WGLQWJ656GUEO/graph.json","events_json":"https://pith.science/api/pith-number/CYKS52THMHZX6WGLQWJ656GUEO/events.json","paper":"https://pith.science/paper/CYKS52TH"},"agent_actions":{"view_html":"https://pith.science/pith/CYKS52THMHZX6WGLQWJ656GUEO","download_json":"https://pith.science/pith/CYKS52THMHZX6WGLQWJ656GUEO.json","view_paper":"https://pith.science/paper/CYKS52TH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.18807&json=true","fetch_graph":"https://pith.science/api/pith-number/CYKS52THMHZX6WGLQWJ656GUEO/graph.json","fetch_events":"https://pith.science/api/pith-number/CYKS52THMHZX6WGLQWJ656GUEO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CYKS52THMHZX6WGLQWJ656GUEO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CYKS52THMHZX6WGLQWJ656GUEO/action/storage_attestation","attest_author":"https://pith.science/pith/CYKS52THMHZX6WGLQWJ656GUEO/action/author_attestation","sign_citation":"https://pith.science/pith/CYKS52THMHZX6WGLQWJ656GUEO/action/citation_signature","submit_replication":"https://pith.science/pith/CYKS52THMHZX6WGLQWJ656GUEO/action/replication_record"}},"created_at":"2026-07-05T11:42:58.629617+00:00","updated_at":"2026-07-05T11:42:58.629617+00:00"}