{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:FL77OHW3LSNC3G5D4H65Z5E7IR","short_pith_number":"pith:FL77OHW3","schema_version":"1.0","canonical_sha256":"2afff71edb5c9a2d9ba3e1fddcf49f446f4120d9d1efd8abeef4a60f74693085","source":{"kind":"arxiv","id":"2509.03594","version":1},"attestation_state":"computed","paper":{"title":"The Optimiser Hidden in Plain Sight: Training with the Loss Landscape's Induced Metric","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","math.OC"],"primary_cat":"cs.LG","authors_text":"Thomas R. Harvey","submitted_at":"2025-09-03T18:00:33Z","abstract_excerpt":"We present a class of novel optimisers for training neural networks that makes use of the Riemannian metric naturally induced when the loss landscape is embedded in higher-dimensional space. This is the same metric that underlies common visualisations of loss landscapes. By taking this geometric perspective literally and using the induced metric, we develop a new optimiser and compare it to existing methods, namely: SGD, Adam, AdamW, and Muon, across a range of tasks and architectures. Empirically, we conclude that this new class of optimisers is highly effective in low dimensional examples, a"},"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":"2509.03594","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-09-03T18:00:33Z","cross_cats_sorted":["cs.AI","math.OC"],"title_canon_sha256":"0615c7af1106ebf479eaf555769e21fa308d5b399121c75b80dafe5aa7512314","abstract_canon_sha256":"2affbaf71abb017d09edfd565cb5972d64424540519fb946e94ee6f87fb46ff6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:04:42.782838Z","signature_b64":"ERzzqwnbOUZl3o6Akul09OKXrnNoM/3ubeJKKbMvMcHytXdV6ku1E4wrFckxfyXMSi06/24bF8lJ2YAmZavgDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2afff71edb5c9a2d9ba3e1fddcf49f446f4120d9d1efd8abeef4a60f74693085","last_reissued_at":"2026-07-05T12:04:42.782334Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:04:42.782334Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Optimiser Hidden in Plain Sight: Training with the Loss Landscape's Induced Metric","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","math.OC"],"primary_cat":"cs.LG","authors_text":"Thomas R. Harvey","submitted_at":"2025-09-03T18:00:33Z","abstract_excerpt":"We present a class of novel optimisers for training neural networks that makes use of the Riemannian metric naturally induced when the loss landscape is embedded in higher-dimensional space. This is the same metric that underlies common visualisations of loss landscapes. By taking this geometric perspective literally and using the induced metric, we develop a new optimiser and compare it to existing methods, namely: SGD, Adam, AdamW, and Muon, across a range of tasks and architectures. Empirically, we conclude that this new class of optimisers is highly effective in low dimensional examples, a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.03594","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/2509.03594/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":"2509.03594","created_at":"2026-07-05T12:04:42.782394+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.03594v1","created_at":"2026-07-05T12:04:42.782394+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.03594","created_at":"2026-07-05T12:04:42.782394+00:00"},{"alias_kind":"pith_short_12","alias_value":"FL77OHW3LSNC","created_at":"2026-07-05T12:04:42.782394+00:00"},{"alias_kind":"pith_short_16","alias_value":"FL77OHW3LSNC3G5D","created_at":"2026-07-05T12:04:42.782394+00:00"},{"alias_kind":"pith_short_8","alias_value":"FL77OHW3","created_at":"2026-07-05T12:04:42.782394+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.05627","citing_title":"Loss-aware state space geometry for quantum variational algorithms","ref_index":67,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FL77OHW3LSNC3G5D4H65Z5E7IR","json":"https://pith.science/pith/FL77OHW3LSNC3G5D4H65Z5E7IR.json","graph_json":"https://pith.science/api/pith-number/FL77OHW3LSNC3G5D4H65Z5E7IR/graph.json","events_json":"https://pith.science/api/pith-number/FL77OHW3LSNC3G5D4H65Z5E7IR/events.json","paper":"https://pith.science/paper/FL77OHW3"},"agent_actions":{"view_html":"https://pith.science/pith/FL77OHW3LSNC3G5D4H65Z5E7IR","download_json":"https://pith.science/pith/FL77OHW3LSNC3G5D4H65Z5E7IR.json","view_paper":"https://pith.science/paper/FL77OHW3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.03594&json=true","fetch_graph":"https://pith.science/api/pith-number/FL77OHW3LSNC3G5D4H65Z5E7IR/graph.json","fetch_events":"https://pith.science/api/pith-number/FL77OHW3LSNC3G5D4H65Z5E7IR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FL77OHW3LSNC3G5D4H65Z5E7IR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FL77OHW3LSNC3G5D4H65Z5E7IR/action/storage_attestation","attest_author":"https://pith.science/pith/FL77OHW3LSNC3G5D4H65Z5E7IR/action/author_attestation","sign_citation":"https://pith.science/pith/FL77OHW3LSNC3G5D4H65Z5E7IR/action/citation_signature","submit_replication":"https://pith.science/pith/FL77OHW3LSNC3G5D4H65Z5E7IR/action/replication_record"}},"created_at":"2026-07-05T12:04:42.782394+00:00","updated_at":"2026-07-05T12:04:42.782394+00:00"}