{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:GWOC47LZKSQ2734QEIOKXVE7TD","short_pith_number":"pith:GWOC47LZ","schema_version":"1.0","canonical_sha256":"359c2e7d7954a1afef90221cabd49f98c19113210374199f18a6a7a163cad532","source":{"kind":"arxiv","id":"1805.06576","version":5},"attestation_state":"computed","paper":{"title":"Mad Max: Affine Spline Insights into Deep Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Randall Balestriero, Richard Baraniuk","submitted_at":"2018-05-17T02:04:54Z","abstract_excerpt":"We build a rigorous bridge between deep networks (DNs) and approximation theory via spline functions and operators. Our key result is that a large class of DNs can be written as a composition of max-affine spline operators (MASOs), which provide a powerful portal through which to view and analyze their inner workings. For instance, conditioned on the input signal, the output of a MASO DN can be written as a simple affine transformation of the input. This implies that a DN constructs a set of signal-dependent, class-specific templates against which the signal is compared via a simple inner prod"},"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":"1805.06576","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-05-17T02:04:54Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"ed71af2a73162feaabc05427a4a30e8bdca0c8236eb71d4a5b2930cefd39c86a","abstract_canon_sha256":"cf5b301da40ebd624be5cb7a34e654ecc7eafe008200b13beb28319224ea92d0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:01:09.353755Z","signature_b64":"v54gX5wbs3euLRlCjRYZUjx/Jyhv8T7x8Kz/86cWETNtxbmbzfbCGi6Pvo69icxspPHEfUX4rdjeaLRRSVF7Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"359c2e7d7954a1afef90221cabd49f98c19113210374199f18a6a7a163cad532","last_reissued_at":"2026-05-18T00:01:09.353178Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:01:09.353178Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mad Max: Affine Spline Insights into Deep Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Randall Balestriero, Richard Baraniuk","submitted_at":"2018-05-17T02:04:54Z","abstract_excerpt":"We build a rigorous bridge between deep networks (DNs) and approximation theory via spline functions and operators. Our key result is that a large class of DNs can be written as a composition of max-affine spline operators (MASOs), which provide a powerful portal through which to view and analyze their inner workings. For instance, conditioned on the input signal, the output of a MASO DN can be written as a simple affine transformation of the input. This implies that a DN constructs a set of signal-dependent, class-specific templates against which the signal is compared via a simple inner prod"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1805.06576","kind":"arxiv","version":5},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"1805.06576","created_at":"2026-05-18T00:01:09.353257+00:00"},{"alias_kind":"arxiv_version","alias_value":"1805.06576v5","created_at":"2026-05-18T00:01:09.353257+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1805.06576","created_at":"2026-05-18T00:01:09.353257+00:00"},{"alias_kind":"pith_short_12","alias_value":"GWOC47LZKSQ2","created_at":"2026-05-18T12:32:25.280505+00:00"},{"alias_kind":"pith_short_16","alias_value":"GWOC47LZKSQ2734Q","created_at":"2026-05-18T12:32:25.280505+00:00"},{"alias_kind":"pith_short_8","alias_value":"GWOC47LZ","created_at":"2026-05-18T12:32:25.280505+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.11962","citing_title":"The Linear Centroids Hypothesis: Features as Directions Learned by Local Experts","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06959","citing_title":"Locally Near Optimal Piecewise Linear Regression in High Dimensions via Difference of Max-Affine Functions","ref_index":159,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11962","citing_title":"The Linear Centroids Hypothesis: Features as Directions Learned by Local Experts","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GWOC47LZKSQ2734QEIOKXVE7TD","json":"https://pith.science/pith/GWOC47LZKSQ2734QEIOKXVE7TD.json","graph_json":"https://pith.science/api/pith-number/GWOC47LZKSQ2734QEIOKXVE7TD/graph.json","events_json":"https://pith.science/api/pith-number/GWOC47LZKSQ2734QEIOKXVE7TD/events.json","paper":"https://pith.science/paper/GWOC47LZ"},"agent_actions":{"view_html":"https://pith.science/pith/GWOC47LZKSQ2734QEIOKXVE7TD","download_json":"https://pith.science/pith/GWOC47LZKSQ2734QEIOKXVE7TD.json","view_paper":"https://pith.science/paper/GWOC47LZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1805.06576&json=true","fetch_graph":"https://pith.science/api/pith-number/GWOC47LZKSQ2734QEIOKXVE7TD/graph.json","fetch_events":"https://pith.science/api/pith-number/GWOC47LZKSQ2734QEIOKXVE7TD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GWOC47LZKSQ2734QEIOKXVE7TD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GWOC47LZKSQ2734QEIOKXVE7TD/action/storage_attestation","attest_author":"https://pith.science/pith/GWOC47LZKSQ2734QEIOKXVE7TD/action/author_attestation","sign_citation":"https://pith.science/pith/GWOC47LZKSQ2734QEIOKXVE7TD/action/citation_signature","submit_replication":"https://pith.science/pith/GWOC47LZKSQ2734QEIOKXVE7TD/action/replication_record"}},"created_at":"2026-05-18T00:01:09.353257+00:00","updated_at":"2026-05-18T00:01:09.353257+00:00"}