{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ZX4YW64ERHGWNDMTZG3XXOK5EB","short_pith_number":"pith:ZX4YW64E","schema_version":"1.0","canonical_sha256":"cdf98b7b8489cd668d93c9b77bb95d2077cbc7364038e39e2a24e3e10f685114","source":{"kind":"arxiv","id":"2606.25971","version":1},"attestation_state":"computed","paper":{"title":"Improving Neural Network Training by Decoupling the Magnitude and Direction of Weight Vectors","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alejandro Hern\\'andez-Cano, Alexander H\\\"agele, Atli Kosson, Martin Jaggi","submitted_at":"2026-06-24T15:40:26Z","abstract_excerpt":"Modern neural network training relies on optimizers such as Adam and Muon which act on each weight matrix as a single object. Yet every weight matrix carries two distinct quantities -- a \\emph{magnitude} and a \\emph{direction} -- and all optimizers stepping in the matrix as a whole couple their dynamics: the directional change from an update depends on the current magnitude, while the magnitude drifts as a byproduct of learning the direction, so neither is governed directly by the learning rate. Typical training therefore leans on surrounding recipes such as weight decay and warmup to keep lea"},"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":"2606.25971","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-24T15:40:26Z","cross_cats_sorted":[],"title_canon_sha256":"a6aff06e6a024bba249f3dc92f28397b61d2e071416a3f361d385d2101f7c7d7","abstract_canon_sha256":"4fb8ba070fa3b8c16a3df6517a93ca78e16e4f8c70b98707b7bcc09141475b3b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-25T01:18:44.723419Z","signature_b64":"cGMq9kipq/Xa9W78j+IqmZkRS9Mc2H+3t0Gy6r6Db+7JgLBOvJkqnSdlgyNssT4BGKvSEfgcrcBmKiUbwDEvDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cdf98b7b8489cd668d93c9b77bb95d2077cbc7364038e39e2a24e3e10f685114","last_reissued_at":"2026-06-25T01:18:44.723013Z","signature_status":"signed_v1","first_computed_at":"2026-06-25T01:18:44.723013Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving Neural Network Training by Decoupling the Magnitude and Direction of Weight Vectors","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alejandro Hern\\'andez-Cano, Alexander H\\\"agele, Atli Kosson, Martin Jaggi","submitted_at":"2026-06-24T15:40:26Z","abstract_excerpt":"Modern neural network training relies on optimizers such as Adam and Muon which act on each weight matrix as a single object. Yet every weight matrix carries two distinct quantities -- a \\emph{magnitude} and a \\emph{direction} -- and all optimizers stepping in the matrix as a whole couple their dynamics: the directional change from an update depends on the current magnitude, while the magnitude drifts as a byproduct of learning the direction, so neither is governed directly by the learning rate. Typical training therefore leans on surrounding recipes such as weight decay and warmup to keep lea"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.25971","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/2606.25971/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":"2606.25971","created_at":"2026-06-25T01:18:44.723075+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.25971v1","created_at":"2026-06-25T01:18:44.723075+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.25971","created_at":"2026-06-25T01:18:44.723075+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZX4YW64ERHGW","created_at":"2026-06-25T01:18:44.723075+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZX4YW64ERHGWNDMT","created_at":"2026-06-25T01:18:44.723075+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZX4YW64E","created_at":"2026-06-25T01:18:44.723075+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZX4YW64ERHGWNDMTZG3XXOK5EB","json":"https://pith.science/pith/ZX4YW64ERHGWNDMTZG3XXOK5EB.json","graph_json":"https://pith.science/api/pith-number/ZX4YW64ERHGWNDMTZG3XXOK5EB/graph.json","events_json":"https://pith.science/api/pith-number/ZX4YW64ERHGWNDMTZG3XXOK5EB/events.json","paper":"https://pith.science/paper/ZX4YW64E"},"agent_actions":{"view_html":"https://pith.science/pith/ZX4YW64ERHGWNDMTZG3XXOK5EB","download_json":"https://pith.science/pith/ZX4YW64ERHGWNDMTZG3XXOK5EB.json","view_paper":"https://pith.science/paper/ZX4YW64E","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.25971&json=true","fetch_graph":"https://pith.science/api/pith-number/ZX4YW64ERHGWNDMTZG3XXOK5EB/graph.json","fetch_events":"https://pith.science/api/pith-number/ZX4YW64ERHGWNDMTZG3XXOK5EB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZX4YW64ERHGWNDMTZG3XXOK5EB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZX4YW64ERHGWNDMTZG3XXOK5EB/action/storage_attestation","attest_author":"https://pith.science/pith/ZX4YW64ERHGWNDMTZG3XXOK5EB/action/author_attestation","sign_citation":"https://pith.science/pith/ZX4YW64ERHGWNDMTZG3XXOK5EB/action/citation_signature","submit_replication":"https://pith.science/pith/ZX4YW64ERHGWNDMTZG3XXOK5EB/action/replication_record"}},"created_at":"2026-06-25T01:18:44.723075+00:00","updated_at":"2026-06-25T01:18:44.723075+00:00"}