{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:75QJ73FWS4DEGRKRDG5DHRC2HZ","short_pith_number":"pith:75QJ73FW","schema_version":"1.0","canonical_sha256":"ff609fecb6970643455119ba33c45a3e73cbeca486e143080415ee5959859bfa","source":{"kind":"arxiv","id":"2507.16178","version":1},"attestation_state":"computed","paper":{"title":"LLM Data Selection and Utilization via Dynamic Bi-level Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Dacheng Tao, Hang Zhou, Kai Han, Kaiqi Huang, Yang Yu, Yehui Tang, Yunhe Wang","submitted_at":"2025-07-22T02:47:12Z","abstract_excerpt":"While large-scale training data is fundamental for developing capable large language models (LLMs), strategically selecting high-quality data has emerged as a critical approach to enhance training efficiency and reduce computational costs. Current data selection methodologies predominantly rely on static, training-agnostic criteria, failing to account for the dynamic model training and data interactions. In this paper, we propose a new Data Weighting Model (DWM) to adjust the weight of selected data within each batch to achieve a dynamic data utilization during LLM training. Specially, to bett"},"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":"2507.16178","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-07-22T02:47:12Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e7653280f85ceae24765501a4f1edc8c498b7f091ff51e68b151d2d195732fb8","abstract_canon_sha256":"e57384890bf26edf728fdd6f210015b7e3a9fd6f1bdbcf81bb14e00eee9e588d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:41:04.638695Z","signature_b64":"qo+lL/0lC8F8qwBQanPqFnKYdrgAMI+f9Aj0DPqSRlwEQjN8XZzv+K+AzRL2O23gwuok6pMYw/qPjejCoj10Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ff609fecb6970643455119ba33c45a3e73cbeca486e143080415ee5959859bfa","last_reissued_at":"2026-07-05T11:41:04.638308Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:41:04.638308Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LLM Data Selection and Utilization via Dynamic Bi-level Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Dacheng Tao, Hang Zhou, Kai Han, Kaiqi Huang, Yang Yu, Yehui Tang, Yunhe Wang","submitted_at":"2025-07-22T02:47:12Z","abstract_excerpt":"While large-scale training data is fundamental for developing capable large language models (LLMs), strategically selecting high-quality data has emerged as a critical approach to enhance training efficiency and reduce computational costs. Current data selection methodologies predominantly rely on static, training-agnostic criteria, failing to account for the dynamic model training and data interactions. In this paper, we propose a new Data Weighting Model (DWM) to adjust the weight of selected data within each batch to achieve a dynamic data utilization during LLM training. Specially, to bett"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.16178","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/2507.16178/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":"2507.16178","created_at":"2026-07-05T11:41:04.638381+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.16178v1","created_at":"2026-07-05T11:41:04.638381+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.16178","created_at":"2026-07-05T11:41:04.638381+00:00"},{"alias_kind":"pith_short_12","alias_value":"75QJ73FWS4DE","created_at":"2026-07-05T11:41:04.638381+00:00"},{"alias_kind":"pith_short_16","alias_value":"75QJ73FWS4DEGRKR","created_at":"2026-07-05T11:41:04.638381+00:00"},{"alias_kind":"pith_short_8","alias_value":"75QJ73FW","created_at":"2026-07-05T11:41:04.638381+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/75QJ73FWS4DEGRKRDG5DHRC2HZ","json":"https://pith.science/pith/75QJ73FWS4DEGRKRDG5DHRC2HZ.json","graph_json":"https://pith.science/api/pith-number/75QJ73FWS4DEGRKRDG5DHRC2HZ/graph.json","events_json":"https://pith.science/api/pith-number/75QJ73FWS4DEGRKRDG5DHRC2HZ/events.json","paper":"https://pith.science/paper/75QJ73FW"},"agent_actions":{"view_html":"https://pith.science/pith/75QJ73FWS4DEGRKRDG5DHRC2HZ","download_json":"https://pith.science/pith/75QJ73FWS4DEGRKRDG5DHRC2HZ.json","view_paper":"https://pith.science/paper/75QJ73FW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.16178&json=true","fetch_graph":"https://pith.science/api/pith-number/75QJ73FWS4DEGRKRDG5DHRC2HZ/graph.json","fetch_events":"https://pith.science/api/pith-number/75QJ73FWS4DEGRKRDG5DHRC2HZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/75QJ73FWS4DEGRKRDG5DHRC2HZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/75QJ73FWS4DEGRKRDG5DHRC2HZ/action/storage_attestation","attest_author":"https://pith.science/pith/75QJ73FWS4DEGRKRDG5DHRC2HZ/action/author_attestation","sign_citation":"https://pith.science/pith/75QJ73FWS4DEGRKRDG5DHRC2HZ/action/citation_signature","submit_replication":"https://pith.science/pith/75QJ73FWS4DEGRKRDG5DHRC2HZ/action/replication_record"}},"created_at":"2026-07-05T11:41:04.638381+00:00","updated_at":"2026-07-05T11:41:04.638381+00:00"}