{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:WFEQ77UAP7CC4IVN67VWQ357YD","short_pith_number":"pith:WFEQ77UA","schema_version":"1.0","canonical_sha256":"b1490ffe807fc42e22adf7eb686fbfc0edfd0d10424b79bf3553d0c0e8612679","source":{"kind":"arxiv","id":"1912.09363","version":3},"attestation_state":"computed","paper":{"title":"Temporal Fusion Transformers for Interpretable Multi-horizon Time Series Forecasting","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Bryan Lim, Nicolas Loeff, Sercan O. Arik, Tomas Pfister","submitted_at":"2019-12-19T16:45:40Z","abstract_excerpt":"Multi-horizon forecasting problems often contain a complex mix of inputs -- including static (i.e. time-invariant) covariates, known future inputs, and other exogenous time series that are only observed historically -- without any prior information on how they interact with the target. While several deep learning models have been proposed for multi-step prediction, they typically comprise black-box models which do not account for the full range of inputs present in common scenarios. In this paper, we introduce the Temporal Fusion Transformer (TFT) -- a novel attention-based architecture which "},"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":"1912.09363","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2019-12-19T16:45:40Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"c47e09d877077edacd47198aa23890bfffcc1b1f1a0051505068045231a3cfd1","abstract_canon_sha256":"32cce22e4fa02fc0b9382efd5f512cf53b1919185b157eac738d62a546d903ad"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:38:16.438450Z","signature_b64":"5GTz6uwA57B5k3l5yglM9gMdpjgbCpu0BOtrL00qC85V+ajnKG9xuCdNc4oYrUmYF5SZXtMJMZFrI9kIaPvhAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b1490ffe807fc42e22adf7eb686fbfc0edfd0d10424b79bf3553d0c0e8612679","last_reissued_at":"2026-07-05T01:38:16.438031Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:38:16.438031Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Temporal Fusion Transformers for Interpretable Multi-horizon Time Series Forecasting","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Bryan Lim, Nicolas Loeff, Sercan O. Arik, Tomas Pfister","submitted_at":"2019-12-19T16:45:40Z","abstract_excerpt":"Multi-horizon forecasting problems often contain a complex mix of inputs -- including static (i.e. time-invariant) covariates, known future inputs, and other exogenous time series that are only observed historically -- without any prior information on how they interact with the target. While several deep learning models have been proposed for multi-step prediction, they typically comprise black-box models which do not account for the full range of inputs present in common scenarios. In this paper, we introduce the Temporal Fusion Transformer (TFT) -- a novel attention-based architecture which "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1912.09363","kind":"arxiv","version":3},"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/1912.09363/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":"1912.09363","created_at":"2026-07-05T01:38:16.438090+00:00"},{"alias_kind":"arxiv_version","alias_value":"1912.09363v3","created_at":"2026-07-05T01:38:16.438090+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1912.09363","created_at":"2026-07-05T01:38:16.438090+00:00"},{"alias_kind":"pith_short_12","alias_value":"WFEQ77UAP7CC","created_at":"2026-07-05T01:38:16.438090+00:00"},{"alias_kind":"pith_short_16","alias_value":"WFEQ77UAP7CC4IVN","created_at":"2026-07-05T01:38:16.438090+00:00"},{"alias_kind":"pith_short_8","alias_value":"WFEQ77UA","created_at":"2026-07-05T01:38:16.438090+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18824","citing_title":"Where Will They Go? Modelling Multimodal Pedestrian Manoeuvres from Ego-centric Videos","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01966","citing_title":"Probabilistic Low-Voltage Peak Load Forecasting with Time Series Foundation Models Evaluated on Application-Oriented Metrics","ref_index":49,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00197","citing_title":"StateFlow: Dual-State Recurrent Modeling for Long-Horizon Time Series Forecasting","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04833","citing_title":"Signed Dual Attention: Capturing Signed Dependencies in Time Series Forecasting","ref_index":12,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WFEQ77UAP7CC4IVN67VWQ357YD","json":"https://pith.science/pith/WFEQ77UAP7CC4IVN67VWQ357YD.json","graph_json":"https://pith.science/api/pith-number/WFEQ77UAP7CC4IVN67VWQ357YD/graph.json","events_json":"https://pith.science/api/pith-number/WFEQ77UAP7CC4IVN67VWQ357YD/events.json","paper":"https://pith.science/paper/WFEQ77UA"},"agent_actions":{"view_html":"https://pith.science/pith/WFEQ77UAP7CC4IVN67VWQ357YD","download_json":"https://pith.science/pith/WFEQ77UAP7CC4IVN67VWQ357YD.json","view_paper":"https://pith.science/paper/WFEQ77UA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1912.09363&json=true","fetch_graph":"https://pith.science/api/pith-number/WFEQ77UAP7CC4IVN67VWQ357YD/graph.json","fetch_events":"https://pith.science/api/pith-number/WFEQ77UAP7CC4IVN67VWQ357YD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WFEQ77UAP7CC4IVN67VWQ357YD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WFEQ77UAP7CC4IVN67VWQ357YD/action/storage_attestation","attest_author":"https://pith.science/pith/WFEQ77UAP7CC4IVN67VWQ357YD/action/author_attestation","sign_citation":"https://pith.science/pith/WFEQ77UAP7CC4IVN67VWQ357YD/action/citation_signature","submit_replication":"https://pith.science/pith/WFEQ77UAP7CC4IVN67VWQ357YD/action/replication_record"}},"created_at":"2026-07-05T01:38:16.438090+00:00","updated_at":"2026-07-05T01:38:16.438090+00:00"}