{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:K6YR7OS5P4BQPIKISPULQHZNKU","short_pith_number":"pith:K6YR7OS5","schema_version":"1.0","canonical_sha256":"57b11fba5d7f0307a14893e8b81f2d5512433fe247e9fbf804a9fd0c4483b613","source":{"kind":"arxiv","id":"2210.04909","version":2},"attestation_state":"computed","paper":{"title":"Meta-Principled Family of Hyperparameter Scaling Strategies","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["hep-th","stat.ML"],"primary_cat":"cs.LG","authors_text":"Sho Yaida","submitted_at":"2022-10-10T18:00:01Z","abstract_excerpt":"In this note, we first derive a one-parameter family of hyperparameter scaling strategies that interpolates between the neural-tangent scaling and mean-field/maximal-update scaling. We then calculate the scalings of dynamical observables -- network outputs, neural tangent kernels, and differentials of neural tangent kernels -- for wide and deep neural networks. These calculations in turn reveal a proper way to scale depth with width such that resultant large-scale models maintain their representation-learning ability. Finally, we observe that various infinite-width limits examined in the liter"},"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":"2210.04909","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-10-10T18:00:01Z","cross_cats_sorted":["hep-th","stat.ML"],"title_canon_sha256":"d1be768dd2271f238135c09dfa54c987f6de4a5a11dc69fd8b8f58ca598f809d","abstract_canon_sha256":"beaa0cfce944e6e0adce19728a7f4028d14713166af192362c761ae8246f37de"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:07:54.832755Z","signature_b64":"cw5ysaGkwqFnTQYtpq3DwsXP2D6NsOPBIMTNPGjK5xnAyiiXstfvTuwMEvMseunELugOzACtdHX2EaOitsLvCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"57b11fba5d7f0307a14893e8b81f2d5512433fe247e9fbf804a9fd0c4483b613","last_reissued_at":"2026-07-05T05:07:54.832365Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:07:54.832365Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Meta-Principled Family of Hyperparameter Scaling Strategies","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["hep-th","stat.ML"],"primary_cat":"cs.LG","authors_text":"Sho Yaida","submitted_at":"2022-10-10T18:00:01Z","abstract_excerpt":"In this note, we first derive a one-parameter family of hyperparameter scaling strategies that interpolates between the neural-tangent scaling and mean-field/maximal-update scaling. We then calculate the scalings of dynamical observables -- network outputs, neural tangent kernels, and differentials of neural tangent kernels -- for wide and deep neural networks. These calculations in turn reveal a proper way to scale depth with width such that resultant large-scale models maintain their representation-learning ability. Finally, we observe that various infinite-width limits examined in the liter"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.04909","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/2210.04909/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":"2210.04909","created_at":"2026-07-05T05:07:54.832421+00:00"},{"alias_kind":"arxiv_version","alias_value":"2210.04909v2","created_at":"2026-07-05T05:07:54.832421+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.04909","created_at":"2026-07-05T05:07:54.832421+00:00"},{"alias_kind":"pith_short_12","alias_value":"K6YR7OS5P4BQ","created_at":"2026-07-05T05:07:54.832421+00:00"},{"alias_kind":"pith_short_16","alias_value":"K6YR7OS5P4BQPIKI","created_at":"2026-07-05T05:07:54.832421+00:00"},{"alias_kind":"pith_short_8","alias_value":"K6YR7OS5","created_at":"2026-07-05T05:07:54.832421+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.21486","citing_title":"Quantifying Hyperparameter Transfer and the Importance of Embedding Layer Learning Rate","ref_index":50,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/K6YR7OS5P4BQPIKISPULQHZNKU","json":"https://pith.science/pith/K6YR7OS5P4BQPIKISPULQHZNKU.json","graph_json":"https://pith.science/api/pith-number/K6YR7OS5P4BQPIKISPULQHZNKU/graph.json","events_json":"https://pith.science/api/pith-number/K6YR7OS5P4BQPIKISPULQHZNKU/events.json","paper":"https://pith.science/paper/K6YR7OS5"},"agent_actions":{"view_html":"https://pith.science/pith/K6YR7OS5P4BQPIKISPULQHZNKU","download_json":"https://pith.science/pith/K6YR7OS5P4BQPIKISPULQHZNKU.json","view_paper":"https://pith.science/paper/K6YR7OS5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2210.04909&json=true","fetch_graph":"https://pith.science/api/pith-number/K6YR7OS5P4BQPIKISPULQHZNKU/graph.json","fetch_events":"https://pith.science/api/pith-number/K6YR7OS5P4BQPIKISPULQHZNKU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/K6YR7OS5P4BQPIKISPULQHZNKU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/K6YR7OS5P4BQPIKISPULQHZNKU/action/storage_attestation","attest_author":"https://pith.science/pith/K6YR7OS5P4BQPIKISPULQHZNKU/action/author_attestation","sign_citation":"https://pith.science/pith/K6YR7OS5P4BQPIKISPULQHZNKU/action/citation_signature","submit_replication":"https://pith.science/pith/K6YR7OS5P4BQPIKISPULQHZNKU/action/replication_record"}},"created_at":"2026-07-05T05:07:54.832421+00:00","updated_at":"2026-07-05T05:07:54.832421+00:00"}