{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:XTULRBSPXNSLEIT764MC4OESOE","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":"79caba003ea1de16804143f5c3e05a8c92c4745c744ce8e2c5717a87e31c4d6b","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-07-28T20:39:16Z","title_canon_sha256":"e67decb3d6255eaa44351b2e1e421540118482578d81d1166db07d0bb978a0bf"},"schema_version":"1.0","source":{"id":"2407.19580","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.19580","created_at":"2026-07-05T11:10:46Z"},{"alias_kind":"arxiv_version","alias_value":"2407.19580v4","created_at":"2026-07-05T11:10:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.19580","created_at":"2026-07-05T11:10:46Z"},{"alias_kind":"pith_short_12","alias_value":"XTULRBSPXNSL","created_at":"2026-07-05T11:10:46Z"},{"alias_kind":"pith_short_16","alias_value":"XTULRBSPXNSLEIT7","created_at":"2026-07-05T11:10:46Z"},{"alias_kind":"pith_short_8","alias_value":"XTULRBSP","created_at":"2026-07-05T11:10:46Z"}],"graph_snapshots":[{"event_id":"sha256:9e3576bfced913660a99639c231c5b378432da72639c30e9e3dff193c3959f5a","target":"graph","created_at":"2026-07-05T11:10:46Z","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/2407.19580/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Training with larger mini-batches improves the convergence rate and can yield superior performance. However, training with large mini-batches becomes prohibitive for Large Language Models (LLMs), due to the large GPU memory requirement. To address this problem, an effective approach is finding small mini-batch coresets that closely match the gradient of larger mini-batches. However, this approach becomes infeasible and ineffective for LLMs, due to the highly imbalanced mixture of sources in language data, use of the Adam optimizer, and the very large gradient dimensionality of LLMs. In this wo","authors_text":"Baharan Mirzasoleiman, Dang Nguyen, Rathul Anand, Wenhan Yang, Yu Yang","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-07-28T20:39:16Z","title":"Mini-batch Coresets for Memory-efficient Language Model Training on Data Mixtures"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.19580","kind":"arxiv","version":4},"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":"sha256:553b7bb3b4b5fbe9a051e073e7d13de348a8efb1a91726317f64b063f50ca235","target":"record","created_at":"2026-07-05T11:10:46Z","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":"79caba003ea1de16804143f5c3e05a8c92c4745c744ce8e2c5717a87e31c4d6b","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-07-28T20:39:16Z","title_canon_sha256":"e67decb3d6255eaa44351b2e1e421540118482578d81d1166db07d0bb978a0bf"},"schema_version":"1.0","source":{"id":"2407.19580","kind":"arxiv","version":4}},"canonical_sha256":"bce8b8864fbb64b2227ff7182e3892710ad68570c587090334416119938bf3c1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"bce8b8864fbb64b2227ff7182e3892710ad68570c587090334416119938bf3c1","first_computed_at":"2026-07-05T11:10:46.896829Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:10:46.896829Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"exHaes5XqVyEkylX1/t8imMue6ZSGrxWQeui9iyVI29tkr8gSj2GTm70gFL4rnRKU+soSAwS81OHdKMoVZG7BA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:10:46.897364Z","signed_message":"canonical_sha256_bytes"},"source_id":"2407.19580","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:553b7bb3b4b5fbe9a051e073e7d13de348a8efb1a91726317f64b063f50ca235","sha256:9e3576bfced913660a99639c231c5b378432da72639c30e9e3dff193c3959f5a"],"state_sha256":"3b51b7047cf89566624337d28c1e6bf7fc0463e05784e8dd057b28e90421962a"}