{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ZQQ4LCJ2R67CXD7UPOVNNN5WAW","short_pith_number":"pith:ZQQ4LCJ2","schema_version":"1.0","canonical_sha256":"cc21c5893a8fbe2b8ff47baad6b7b605a41b0b7e1dad24cff6506bff5ed8eaa2","source":{"kind":"arxiv","id":"2501.01045","version":4},"attestation_state":"computed","paper":{"title":"ZeroFlow: Overcoming Catastrophic Forgetting is Easier than You Think","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Dan Zhang, Didi Zhu, Hangjie Yuan, Jie Tang, Tao Feng, Wei Li, Wendi Zheng","submitted_at":"2025-01-02T04:10:17Z","abstract_excerpt":"Backpropagation provides a generalized configuration for overcoming catastrophic forgetting. Optimizers such as SGD and Adam are commonly used for weight updates in continual learning and continual pre-training. However, access to gradient information is not always feasible in practice due to black-box APIs, hardware constraints, or non-differentiable systems, a challenge we refer to as the gradient bans. To bridge this gap, we introduce ZeroFlow, the first benchmark designed to evaluate gradient-free optimization algorithms for overcoming forgetting. ZeroFlow examines a suite of forward pass-"},"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":"2501.01045","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-01-02T04:10:17Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"3a39a5dd01cdcbd8743fd98f2fa8fbbb618c5e86e7dbe6a3d54079a51f629e6e","abstract_canon_sha256":"3e1a4c054d07990078d142557d5f719d14013cf93fe51a1d360a5362ff9b695c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:16:53.842085Z","signature_b64":"pqRQFcLWGdZpsr0a18CTdcn2jE+HJK8XZxxLILuSFIQOVFGjFA1AApva/6uWXlYpcZeQz93I2HMCQp23vp7vAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cc21c5893a8fbe2b8ff47baad6b7b605a41b0b7e1dad24cff6506bff5ed8eaa2","last_reissued_at":"2026-07-05T11:16:53.841575Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:16:53.841575Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ZeroFlow: Overcoming Catastrophic Forgetting is Easier than You Think","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Dan Zhang, Didi Zhu, Hangjie Yuan, Jie Tang, Tao Feng, Wei Li, Wendi Zheng","submitted_at":"2025-01-02T04:10:17Z","abstract_excerpt":"Backpropagation provides a generalized configuration for overcoming catastrophic forgetting. Optimizers such as SGD and Adam are commonly used for weight updates in continual learning and continual pre-training. However, access to gradient information is not always feasible in practice due to black-box APIs, hardware constraints, or non-differentiable systems, a challenge we refer to as the gradient bans. To bridge this gap, we introduce ZeroFlow, the first benchmark designed to evaluate gradient-free optimization algorithms for overcoming forgetting. ZeroFlow examines a suite of forward pass-"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.01045","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/2501.01045/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":"2501.01045","created_at":"2026-07-05T11:16:53.841637+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.01045v4","created_at":"2026-07-05T11:16:53.841637+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.01045","created_at":"2026-07-05T11:16:53.841637+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZQQ4LCJ2R67C","created_at":"2026-07-05T11:16:53.841637+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZQQ4LCJ2R67CXD7U","created_at":"2026-07-05T11:16:53.841637+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZQQ4LCJ2","created_at":"2026-07-05T11:16:53.841637+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.18860","citing_title":"C-Flat++: Towards a More Efficient and Powerful Framework for Continual Learning","ref_index":58,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZQQ4LCJ2R67CXD7UPOVNNN5WAW","json":"https://pith.science/pith/ZQQ4LCJ2R67CXD7UPOVNNN5WAW.json","graph_json":"https://pith.science/api/pith-number/ZQQ4LCJ2R67CXD7UPOVNNN5WAW/graph.json","events_json":"https://pith.science/api/pith-number/ZQQ4LCJ2R67CXD7UPOVNNN5WAW/events.json","paper":"https://pith.science/paper/ZQQ4LCJ2"},"agent_actions":{"view_html":"https://pith.science/pith/ZQQ4LCJ2R67CXD7UPOVNNN5WAW","download_json":"https://pith.science/pith/ZQQ4LCJ2R67CXD7UPOVNNN5WAW.json","view_paper":"https://pith.science/paper/ZQQ4LCJ2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.01045&json=true","fetch_graph":"https://pith.science/api/pith-number/ZQQ4LCJ2R67CXD7UPOVNNN5WAW/graph.json","fetch_events":"https://pith.science/api/pith-number/ZQQ4LCJ2R67CXD7UPOVNNN5WAW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZQQ4LCJ2R67CXD7UPOVNNN5WAW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZQQ4LCJ2R67CXD7UPOVNNN5WAW/action/storage_attestation","attest_author":"https://pith.science/pith/ZQQ4LCJ2R67CXD7UPOVNNN5WAW/action/author_attestation","sign_citation":"https://pith.science/pith/ZQQ4LCJ2R67CXD7UPOVNNN5WAW/action/citation_signature","submit_replication":"https://pith.science/pith/ZQQ4LCJ2R67CXD7UPOVNNN5WAW/action/replication_record"}},"created_at":"2026-07-05T11:16:53.841637+00:00","updated_at":"2026-07-05T11:16:53.841637+00:00"}