{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BKC5MG6KBN5Q52UWJD2WEWIRFF","short_pith_number":"pith:BKC5MG6K","schema_version":"1.0","canonical_sha256":"0a85d61bca0b7b0eea9648f56259112962e6f37f3dfa5043a09e4fd1838943fb","source":{"kind":"arxiv","id":"2506.11029","version":1},"attestation_state":"computed","paper":{"title":"Output Scaling: YingLong-Delayed Chain of Thought in a Large Pretrained Time Series Forecasting Model","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Bolin Ding, Jingren Zhou, Jinyang Gao, Tian Zhou, Xue Wang","submitted_at":"2025-05-20T14:31:06Z","abstract_excerpt":"We present a joint forecasting framework for time series prediction that contrasts with traditional direct or recursive methods. This framework achieves state-of-the-art performance for our designed foundation model, YingLong, and reveals a novel scaling effect: longer outputs significantly enhance model accuracy due to delayed chain-of-thought reasoning in our non-causal approach. YingLong is a non-causal, bidirectional attention encoder-only transformer trained through masked token recovery, aligning more effectively with language understanding tasks than with generation tasks. Additionally,"},"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":"2506.11029","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-20T14:31:06Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7eb7509a20ca59b0468b836f8fbe079b415d68ec0d0e69748a608d4b61a5d7c8","abstract_canon_sha256":"780b3fb44922698dca260c201688ff707ca399da3d19eb4cabaaba1c8fe6e9b5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:20:58.171918Z","signature_b64":"w52Aij3I2rwN93va0FSpLC+sqvmjCkrcZcYSTkZXA7ybYVccOyz7qnq2oE8ITCbPzCyKcsZXtuP4hydt0x1+DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0a85d61bca0b7b0eea9648f56259112962e6f37f3dfa5043a09e4fd1838943fb","last_reissued_at":"2026-07-05T11:20:58.171279Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:20:58.171279Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Output Scaling: YingLong-Delayed Chain of Thought in a Large Pretrained Time Series Forecasting Model","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Bolin Ding, Jingren Zhou, Jinyang Gao, Tian Zhou, Xue Wang","submitted_at":"2025-05-20T14:31:06Z","abstract_excerpt":"We present a joint forecasting framework for time series prediction that contrasts with traditional direct or recursive methods. This framework achieves state-of-the-art performance for our designed foundation model, YingLong, and reveals a novel scaling effect: longer outputs significantly enhance model accuracy due to delayed chain-of-thought reasoning in our non-causal approach. YingLong is a non-causal, bidirectional attention encoder-only transformer trained through masked token recovery, aligning more effectively with language understanding tasks than with generation tasks. Additionally,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.11029","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/2506.11029/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":"2506.11029","created_at":"2026-07-05T11:20:58.171376+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.11029v1","created_at":"2026-07-05T11:20:58.171376+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.11029","created_at":"2026-07-05T11:20:58.171376+00:00"},{"alias_kind":"pith_short_12","alias_value":"BKC5MG6KBN5Q","created_at":"2026-07-05T11:20:58.171376+00:00"},{"alias_kind":"pith_short_16","alias_value":"BKC5MG6KBN5Q52UW","created_at":"2026-07-05T11:20:58.171376+00:00"},{"alias_kind":"pith_short_8","alias_value":"BKC5MG6K","created_at":"2026-07-05T11:20:58.171376+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"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":66,"is_internal_anchor":false},{"citing_arxiv_id":"2606.10798","citing_title":"CITRAS-FM: Tiny Time Series Foundation Model for Covariate-Informed Zero-Shot Forecasting","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01204","citing_title":"TiRex-2: Generalizing TiRex to Multivariate Data and Streaming","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09861","citing_title":"Time Series as Language: A Universal Tokenizer for General-Purpose Time Series Foundation Models","ref_index":47,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10544","citing_title":"WaveMoE: A Wavelet-Enhanced Mixture-of-Experts Foundation Model for Time Series Forecasting","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BKC5MG6KBN5Q52UWJD2WEWIRFF","json":"https://pith.science/pith/BKC5MG6KBN5Q52UWJD2WEWIRFF.json","graph_json":"https://pith.science/api/pith-number/BKC5MG6KBN5Q52UWJD2WEWIRFF/graph.json","events_json":"https://pith.science/api/pith-number/BKC5MG6KBN5Q52UWJD2WEWIRFF/events.json","paper":"https://pith.science/paper/BKC5MG6K"},"agent_actions":{"view_html":"https://pith.science/pith/BKC5MG6KBN5Q52UWJD2WEWIRFF","download_json":"https://pith.science/pith/BKC5MG6KBN5Q52UWJD2WEWIRFF.json","view_paper":"https://pith.science/paper/BKC5MG6K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.11029&json=true","fetch_graph":"https://pith.science/api/pith-number/BKC5MG6KBN5Q52UWJD2WEWIRFF/graph.json","fetch_events":"https://pith.science/api/pith-number/BKC5MG6KBN5Q52UWJD2WEWIRFF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BKC5MG6KBN5Q52UWJD2WEWIRFF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BKC5MG6KBN5Q52UWJD2WEWIRFF/action/storage_attestation","attest_author":"https://pith.science/pith/BKC5MG6KBN5Q52UWJD2WEWIRFF/action/author_attestation","sign_citation":"https://pith.science/pith/BKC5MG6KBN5Q52UWJD2WEWIRFF/action/citation_signature","submit_replication":"https://pith.science/pith/BKC5MG6KBN5Q52UWJD2WEWIRFF/action/replication_record"}},"created_at":"2026-07-05T11:20:58.171376+00:00","updated_at":"2026-07-05T11:20:58.171376+00:00"}