{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:XFQLQ7NPPJ7MV4QT64R6V2CRQS","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":"b8ea47cb5144209052ff0f3ce488e268ac054ccffcfad6e4d55f2bd984a848f4","cross_cats_sorted":["cs.AI","cs.CL","cs.IR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-10-15T23:23:27Z","title_canon_sha256":"1f96aeb1c9af94eeec31d735cd9aa41c85a3992a0bf1b3b9e200227c39239a44"},"schema_version":"1.0","source":{"id":"2310.09983","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.09983","created_at":"2026-07-05T07:01:13Z"},{"alias_kind":"arxiv_version","alias_value":"2310.09983v1","created_at":"2026-07-05T07:01:13Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.09983","created_at":"2026-07-05T07:01:13Z"},{"alias_kind":"pith_short_12","alias_value":"XFQLQ7NPPJ7M","created_at":"2026-07-05T07:01:13Z"},{"alias_kind":"pith_short_16","alias_value":"XFQLQ7NPPJ7MV4QT","created_at":"2026-07-05T07:01:13Z"},{"alias_kind":"pith_short_8","alias_value":"XFQLQ7NP","created_at":"2026-07-05T07:01:13Z"}],"graph_snapshots":[{"event_id":"sha256:6f54149047cfd40e9c01acbe9d8698a506cfc6a4abe1d79ec612b08eaa89f020","target":"graph","created_at":"2026-07-05T07:01:13Z","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/2310.09983/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We study data distillation for auto-regressive machine learning tasks, where the input and output have a strict left-to-right causal structure. More specifically, we propose Farzi, which summarizes an event sequence dataset into a small number of synthetic sequences -- Farzi Data -- which are optimized to maintain (if not improve) model performance compared to training on the full dataset. Under the hood, Farzi conducts memory-efficient data distillation by (i) deriving efficient reverse-mode differentiation of the Adam optimizer by leveraging Hessian-Vector Products; and (ii) factorizing the ","authors_text":"Derek Zhiyuan Cheng, Jianmo Ni, Julian McAuley, Noveen Sachdeva, Wang-Cheng Kang, Zexue He","cross_cats":["cs.AI","cs.CL","cs.IR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-10-15T23:23:27Z","title":"Farzi Data: Autoregressive Data Distillation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.09983","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:af64afdac5a464b1d338988d70e385b37f6fd08aedb35a773aa10a28ce0467d6","target":"record","created_at":"2026-07-05T07:01:13Z","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":"b8ea47cb5144209052ff0f3ce488e268ac054ccffcfad6e4d55f2bd984a848f4","cross_cats_sorted":["cs.AI","cs.CL","cs.IR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-10-15T23:23:27Z","title_canon_sha256":"1f96aeb1c9af94eeec31d735cd9aa41c85a3992a0bf1b3b9e200227c39239a44"},"schema_version":"1.0","source":{"id":"2310.09983","kind":"arxiv","version":1}},"canonical_sha256":"b960b87daf7a7ecaf213f723eae85184a55bf2b2dad04808ae2317df4768d80b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b960b87daf7a7ecaf213f723eae85184a55bf2b2dad04808ae2317df4768d80b","first_computed_at":"2026-07-05T07:01:13.344912Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:01:13.344912Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"FdRVeHJVZvcx5eYmIEad+0f4DRCsI1HVFNRPMuxad9BSFE5inOokyLgQ5ILlw3za8FpCsTPWRn42SnL4xp5HCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T07:01:13.345375Z","signed_message":"canonical_sha256_bytes"},"source_id":"2310.09983","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:af64afdac5a464b1d338988d70e385b37f6fd08aedb35a773aa10a28ce0467d6","sha256:6f54149047cfd40e9c01acbe9d8698a506cfc6a4abe1d79ec612b08eaa89f020"],"state_sha256":"f773a318a0f9c9f5a5119f2c58b0dd08a1cc871af17555a76f8c8e58c8a9fb3b"}