{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:SL2DV3FCPXTXCHNPK24ADEHF6M","short_pith_number":"pith:SL2DV3FC","schema_version":"1.0","canonical_sha256":"92f43aeca27de7711daf56b80190e5f30b2d9e4b85dbc91c1393fd5b7b0ab439","source":{"kind":"arxiv","id":"2508.01474","version":1},"attestation_state":"computed","paper":{"title":"HT-Transformer: Event Sequences Classification by Accumulating Prefix Information with History Tokens","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Andrey Savchenko, Ivan Karpukhin","submitted_at":"2025-08-02T19:50:58Z","abstract_excerpt":"Deep learning has achieved remarkable success in modeling sequential data, including event sequences, temporal point processes, and irregular time series. Recently, transformers have largely replaced recurrent networks in these tasks. However, transformers often underperform RNNs in classification tasks where the objective is to predict future targets. The reason behind this performance gap remains largely unexplored. In this paper, we identify a key limitation of transformers: the absence of a single state vector that provides a compact and effective representation of the entire sequence. Add"},"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":"2508.01474","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-02T19:50:58Z","cross_cats_sorted":[],"title_canon_sha256":"eb728d7118697d612c400388947d6d02d38ada79dcb6dfcc00982f4040484049","abstract_canon_sha256":"bc662cee796807b255b682ead4e8565ea87473252f67aba0b8c1cf07785cc1b5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:47:39.723340Z","signature_b64":"10RROI/v1wKypHZ1iIMClyILIHFgatCroHYsGEss6GXkYFrW0yf5M/j6wamHNrUNUp6yQJMYFotu/c/yESukCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"92f43aeca27de7711daf56b80190e5f30b2d9e4b85dbc91c1393fd5b7b0ab439","last_reissued_at":"2026-07-05T11:47:39.722869Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:47:39.722869Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HT-Transformer: Event Sequences Classification by Accumulating Prefix Information with History Tokens","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Andrey Savchenko, Ivan Karpukhin","submitted_at":"2025-08-02T19:50:58Z","abstract_excerpt":"Deep learning has achieved remarkable success in modeling sequential data, including event sequences, temporal point processes, and irregular time series. Recently, transformers have largely replaced recurrent networks in these tasks. However, transformers often underperform RNNs in classification tasks where the objective is to predict future targets. The reason behind this performance gap remains largely unexplored. In this paper, we identify a key limitation of transformers: the absence of a single state vector that provides a compact and effective representation of the entire sequence. Add"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.01474","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/2508.01474/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":"2508.01474","created_at":"2026-07-05T11:47:39.722926+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.01474v1","created_at":"2026-07-05T11:47:39.722926+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.01474","created_at":"2026-07-05T11:47:39.722926+00:00"},{"alias_kind":"pith_short_12","alias_value":"SL2DV3FCPXTX","created_at":"2026-07-05T11:47:39.722926+00:00"},{"alias_kind":"pith_short_16","alias_value":"SL2DV3FCPXTXCHNP","created_at":"2026-07-05T11:47:39.722926+00:00"},{"alias_kind":"pith_short_8","alias_value":"SL2DV3FC","created_at":"2026-07-05T11:47:39.722926+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.07756","citing_title":"When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SL2DV3FCPXTXCHNPK24ADEHF6M","json":"https://pith.science/pith/SL2DV3FCPXTXCHNPK24ADEHF6M.json","graph_json":"https://pith.science/api/pith-number/SL2DV3FCPXTXCHNPK24ADEHF6M/graph.json","events_json":"https://pith.science/api/pith-number/SL2DV3FCPXTXCHNPK24ADEHF6M/events.json","paper":"https://pith.science/paper/SL2DV3FC"},"agent_actions":{"view_html":"https://pith.science/pith/SL2DV3FCPXTXCHNPK24ADEHF6M","download_json":"https://pith.science/pith/SL2DV3FCPXTXCHNPK24ADEHF6M.json","view_paper":"https://pith.science/paper/SL2DV3FC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.01474&json=true","fetch_graph":"https://pith.science/api/pith-number/SL2DV3FCPXTXCHNPK24ADEHF6M/graph.json","fetch_events":"https://pith.science/api/pith-number/SL2DV3FCPXTXCHNPK24ADEHF6M/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SL2DV3FCPXTXCHNPK24ADEHF6M/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SL2DV3FCPXTXCHNPK24ADEHF6M/action/storage_attestation","attest_author":"https://pith.science/pith/SL2DV3FCPXTXCHNPK24ADEHF6M/action/author_attestation","sign_citation":"https://pith.science/pith/SL2DV3FCPXTXCHNPK24ADEHF6M/action/citation_signature","submit_replication":"https://pith.science/pith/SL2DV3FCPXTXCHNPK24ADEHF6M/action/replication_record"}},"created_at":"2026-07-05T11:47:39.722926+00:00","updated_at":"2026-07-05T11:47:39.722926+00:00"}