{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:74IO3BEJEWBEKUEBKMEHUKOFPD","short_pith_number":"pith:74IO3BEJ","schema_version":"1.0","canonical_sha256":"ff10ed8489258245508153087a29c578db0edc5b8278ba170817ad13814fdb37","source":{"kind":"arxiv","id":"2112.00980","version":4},"attestation_state":"computed","paper":{"title":"Trap of Feature Diversity in the Learning of MLPs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Dongrui Liu, Huiqi Deng, Jie Ren, Kangrui Wang, Quanshi Zhang, Shaobo Wang, Sheng Yin","submitted_at":"2021-12-02T04:42:26Z","abstract_excerpt":"In this paper, we focus on a typical two-phase phenomenon in the learning of multi-layer perceptrons (MLPs), and we aim to explain the reason for the decrease of feature diversity in the first phase. Specifically, people find that, in the training of MLPs, the training loss does not decrease significantly until the second phase. To this end, we further explore the reason why the diversity of features over different samples keeps decreasing in the first phase, which hurts the optimization of MLPs. We explain such a phenomenon in terms of the learning dynamics of MLPs. Furthermore, we theoretica"},"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":"2112.00980","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-12-02T04:42:26Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7aff945b61232bc59cfd66ae51be83849696e60d266b24e7a9e20b35a7d73378","abstract_canon_sha256":"7e34f227e42f27ce977a2b83a6a14d8ed28fa915cc67d88933c4e747c9399d04"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:28:08.804955Z","signature_b64":"U8uHQNSrqUC9swhV0kfBBPuwz+lNbm1kH9xvLOHJAMXdrCrQGxlPhKoAYXgud85R32fto4CkGdSC32FVAhEDCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ff10ed8489258245508153087a29c578db0edc5b8278ba170817ad13814fdb37","last_reissued_at":"2026-07-05T04:28:08.804425Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:28:08.804425Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Trap of Feature Diversity in the Learning of MLPs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Dongrui Liu, Huiqi Deng, Jie Ren, Kangrui Wang, Quanshi Zhang, Shaobo Wang, Sheng Yin","submitted_at":"2021-12-02T04:42:26Z","abstract_excerpt":"In this paper, we focus on a typical two-phase phenomenon in the learning of multi-layer perceptrons (MLPs), and we aim to explain the reason for the decrease of feature diversity in the first phase. Specifically, people find that, in the training of MLPs, the training loss does not decrease significantly until the second phase. To this end, we further explore the reason why the diversity of features over different samples keeps decreasing in the first phase, which hurts the optimization of MLPs. We explain such a phenomenon in terms of the learning dynamics of MLPs. Furthermore, we theoretica"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2112.00980","kind":"arxiv","version":4},"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/2112.00980/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":"2112.00980","created_at":"2026-07-05T04:28:08.804485+00:00"},{"alias_kind":"arxiv_version","alias_value":"2112.00980v4","created_at":"2026-07-05T04:28:08.804485+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2112.00980","created_at":"2026-07-05T04:28:08.804485+00:00"},{"alias_kind":"pith_short_12","alias_value":"74IO3BEJEWBE","created_at":"2026-07-05T04:28:08.804485+00:00"},{"alias_kind":"pith_short_16","alias_value":"74IO3BEJEWBEKUEB","created_at":"2026-07-05T04:28:08.804485+00:00"},{"alias_kind":"pith_short_8","alias_value":"74IO3BEJ","created_at":"2026-07-05T04:28:08.804485+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/74IO3BEJEWBEKUEBKMEHUKOFPD","json":"https://pith.science/pith/74IO3BEJEWBEKUEBKMEHUKOFPD.json","graph_json":"https://pith.science/api/pith-number/74IO3BEJEWBEKUEBKMEHUKOFPD/graph.json","events_json":"https://pith.science/api/pith-number/74IO3BEJEWBEKUEBKMEHUKOFPD/events.json","paper":"https://pith.science/paper/74IO3BEJ"},"agent_actions":{"view_html":"https://pith.science/pith/74IO3BEJEWBEKUEBKMEHUKOFPD","download_json":"https://pith.science/pith/74IO3BEJEWBEKUEBKMEHUKOFPD.json","view_paper":"https://pith.science/paper/74IO3BEJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2112.00980&json=true","fetch_graph":"https://pith.science/api/pith-number/74IO3BEJEWBEKUEBKMEHUKOFPD/graph.json","fetch_events":"https://pith.science/api/pith-number/74IO3BEJEWBEKUEBKMEHUKOFPD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/74IO3BEJEWBEKUEBKMEHUKOFPD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/74IO3BEJEWBEKUEBKMEHUKOFPD/action/storage_attestation","attest_author":"https://pith.science/pith/74IO3BEJEWBEKUEBKMEHUKOFPD/action/author_attestation","sign_citation":"https://pith.science/pith/74IO3BEJEWBEKUEBKMEHUKOFPD/action/citation_signature","submit_replication":"https://pith.science/pith/74IO3BEJEWBEKUEBKMEHUKOFPD/action/replication_record"}},"created_at":"2026-07-05T04:28:08.804485+00:00","updated_at":"2026-07-05T04:28:08.804485+00:00"}