{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:PZD5BABZ2FQB6WUFEGMJAAWC6L","short_pith_number":"pith:PZD5BABZ","canonical_record":{"source":{"id":"2410.23819","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-31T11:04:07Z","cross_cats_sorted":[],"title_canon_sha256":"6505807be5b6fd450bd7159ddfa7f86b43cbae20f4efedbfda389b9798ccf7a2","abstract_canon_sha256":"8f2c56043b7a72398bb5de25acfcbd7f074e46460e395746b10abdf62a82c2d1"},"schema_version":"1.0"},"canonical_sha256":"7e47d08039d1601f5a8521989002c2f2ff4bdaba37bc3421f6c5c0c5e4ceba13","source":{"kind":"arxiv","id":"2410.23819","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.23819","created_at":"2026-07-05T09:29:05Z"},{"alias_kind":"arxiv_version","alias_value":"2410.23819v1","created_at":"2026-07-05T09:29:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.23819","created_at":"2026-07-05T09:29:05Z"},{"alias_kind":"pith_short_12","alias_value":"PZD5BABZ2FQB","created_at":"2026-07-05T09:29:05Z"},{"alias_kind":"pith_short_16","alias_value":"PZD5BABZ2FQB6WUF","created_at":"2026-07-05T09:29:05Z"},{"alias_kind":"pith_short_8","alias_value":"PZD5BABZ","created_at":"2026-07-05T09:29:05Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:PZD5BABZ2FQB6WUFEGMJAAWC6L","target":"record","payload":{"canonical_record":{"source":{"id":"2410.23819","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-31T11:04:07Z","cross_cats_sorted":[],"title_canon_sha256":"6505807be5b6fd450bd7159ddfa7f86b43cbae20f4efedbfda389b9798ccf7a2","abstract_canon_sha256":"8f2c56043b7a72398bb5de25acfcbd7f074e46460e395746b10abdf62a82c2d1"},"schema_version":"1.0"},"canonical_sha256":"7e47d08039d1601f5a8521989002c2f2ff4bdaba37bc3421f6c5c0c5e4ceba13","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:29:05.712091Z","signature_b64":"52rmCWOv3IjtFDSyhI5+mmzSLKeOWUBfIYyBl9n381EdSEQDyhWYFfDaDsbPE+FmXUuBRbOAVPLUOoB2PDysCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7e47d08039d1601f5a8521989002c2f2ff4bdaba37bc3421f6c5c0c5e4ceba13","last_reissued_at":"2026-07-05T09:29:05.711599Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:29:05.711599Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.23819","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:29:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7gAtz/qpCPY/5LROnitpZ4dxbAr+d/OQqvdHKWng/H49CXaycIOdAw39WfHSZP+j4+KcvUJbnA0El4NY+EQ+CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T10:11:37.352725Z"},"content_sha256":"4bdc5b274d36bc5c869f00b46bb70cd7615b089ca4485d3cec7bdf6df97f0afc","schema_version":"1.0","event_id":"sha256:4bdc5b274d36bc5c869f00b46bb70cd7615b089ca4485d3cec7bdf6df97f0afc"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:PZD5BABZ2FQB6WUFEGMJAAWC6L","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Weight decay induces low-rank attention layers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Johannes von Oswald, Seijin Kobayashi, Yassir Akram","submitted_at":"2024-10-31T11:04:07Z","abstract_excerpt":"The effect of regularizers such as weight decay when training deep neural networks is not well understood. We study the influence of weight decay as well as $L2$-regularization when training neural network models in which parameter matrices interact multiplicatively. This combination is of particular interest as this parametrization is common in attention layers, the workhorse of transformers. Here, key-query, as well as value-projection parameter matrices, are multiplied directly with each other: $W_K^TW_Q$ and $PW_V$. We extend previous results and show on one hand that any local minimum of "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.23819","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/2410.23819/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:29:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uPg3hc7vYB5eofo8IMNKBGC7mdbbmaGOJu07g7jaSangEbysbdaycizxiY0PygemPT6VPxiCu0cOkJLhVzm4AQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T10:11:37.353240Z"},"content_sha256":"0353600e2f0ac3aa3ee3f0388dc8a3fea89f11217cf89e3298bd6f48e8f2b191","schema_version":"1.0","event_id":"sha256:0353600e2f0ac3aa3ee3f0388dc8a3fea89f11217cf89e3298bd6f48e8f2b191"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/PZD5BABZ2FQB6WUFEGMJAAWC6L/bundle.json","state_url":"https://pith.science/pith/PZD5BABZ2FQB6WUFEGMJAAWC6L/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/PZD5BABZ2FQB6WUFEGMJAAWC6L/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-10T10:11:37Z","links":{"resolver":"https://pith.science/pith/PZD5BABZ2FQB6WUFEGMJAAWC6L","bundle":"https://pith.science/pith/PZD5BABZ2FQB6WUFEGMJAAWC6L/bundle.json","state":"https://pith.science/pith/PZD5BABZ2FQB6WUFEGMJAAWC6L/state.json","well_known_bundle":"https://pith.science/.well-known/pith/PZD5BABZ2FQB6WUFEGMJAAWC6L/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:PZD5BABZ2FQB6WUFEGMJAAWC6L","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"8f2c56043b7a72398bb5de25acfcbd7f074e46460e395746b10abdf62a82c2d1","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-31T11:04:07Z","title_canon_sha256":"6505807be5b6fd450bd7159ddfa7f86b43cbae20f4efedbfda389b9798ccf7a2"},"schema_version":"1.0","source":{"id":"2410.23819","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.23819","created_at":"2026-07-05T09:29:05Z"},{"alias_kind":"arxiv_version","alias_value":"2410.23819v1","created_at":"2026-07-05T09:29:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.23819","created_at":"2026-07-05T09:29:05Z"},{"alias_kind":"pith_short_12","alias_value":"PZD5BABZ2FQB","created_at":"2026-07-05T09:29:05Z"},{"alias_kind":"pith_short_16","alias_value":"PZD5BABZ2FQB6WUF","created_at":"2026-07-05T09:29:05Z"},{"alias_kind":"pith_short_8","alias_value":"PZD5BABZ","created_at":"2026-07-05T09:29:05Z"}],"graph_snapshots":[{"event_id":"sha256:0353600e2f0ac3aa3ee3f0388dc8a3fea89f11217cf89e3298bd6f48e8f2b191","target":"graph","created_at":"2026-07-05T09:29:05Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2410.23819/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The effect of regularizers such as weight decay when training deep neural networks is not well understood. We study the influence of weight decay as well as $L2$-regularization when training neural network models in which parameter matrices interact multiplicatively. This combination is of particular interest as this parametrization is common in attention layers, the workhorse of transformers. Here, key-query, as well as value-projection parameter matrices, are multiplied directly with each other: $W_K^TW_Q$ and $PW_V$. We extend previous results and show on one hand that any local minimum of ","authors_text":"Johannes von Oswald, Seijin Kobayashi, Yassir Akram","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-31T11:04:07Z","title":"Weight decay induces low-rank attention layers"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.23819","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:4bdc5b274d36bc5c869f00b46bb70cd7615b089ca4485d3cec7bdf6df97f0afc","target":"record","created_at":"2026-07-05T09:29:05Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"8f2c56043b7a72398bb5de25acfcbd7f074e46460e395746b10abdf62a82c2d1","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-31T11:04:07Z","title_canon_sha256":"6505807be5b6fd450bd7159ddfa7f86b43cbae20f4efedbfda389b9798ccf7a2"},"schema_version":"1.0","source":{"id":"2410.23819","kind":"arxiv","version":1}},"canonical_sha256":"7e47d08039d1601f5a8521989002c2f2ff4bdaba37bc3421f6c5c0c5e4ceba13","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7e47d08039d1601f5a8521989002c2f2ff4bdaba37bc3421f6c5c0c5e4ceba13","first_computed_at":"2026-07-05T09:29:05.711599Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:29:05.711599Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"52rmCWOv3IjtFDSyhI5+mmzSLKeOWUBfIYyBl9n381EdSEQDyhWYFfDaDsbPE+FmXUuBRbOAVPLUOoB2PDysCg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:29:05.712091Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.23819","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4bdc5b274d36bc5c869f00b46bb70cd7615b089ca4485d3cec7bdf6df97f0afc","sha256:0353600e2f0ac3aa3ee3f0388dc8a3fea89f11217cf89e3298bd6f48e8f2b191"],"state_sha256":"b7f766fd5a74f563f4620845d699af507ef351be1256804413e076905f0482fa"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"UF0n434kkx6Vwc+tYZMIk1eXG05GB7B9O5oC+BuABp6UniDMRoPT9dPQbYFvyte1TG3KOmeYxJ8zJlscI4wrBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T10:11:37.357353Z","bundle_sha256":"8fea3f8ba27e67760bff94118716dfbc6ae24b8c9723b3ed94d9c852f6bc4635"}}