{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:DMUIVTJVBFYOSVIZHZH443YPG6","short_pith_number":"pith:DMUIVTJV","schema_version":"1.0","canonical_sha256":"1b288acd350970e955193e4fce6f0f37bb785b217462bafe1c85711d592a442a","source":{"kind":"arxiv","id":"2310.01728","version":2},"attestation_state":"computed","paper":{"title":"Time-LLM: Time Series Forecasting by Reprogramming Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Reprogramming time series inputs with text prototypes lets frozen large language models generate accurate forecasts without retraining.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"James Y. Zhang, Lintao Ma, Ming Jin, Pin-Yu Chen, Qingsong Wen, Shirui Pan, Shiyu Wang, Xiaoming Shi, Yuan-Fang Li, Yuxuan Liang, Zhixuan Chu","submitted_at":"2023-10-03T01:31:25Z","abstract_excerpt":"Time series forecasting holds significant importance in many real-world dynamic systems and has been extensively studied. Unlike natural language process (NLP) and computer vision (CV), where a single large model can tackle multiple tasks, models for time series forecasting are often specialized, necessitating distinct designs for different tasks and applications. While pre-trained foundation models have made impressive strides in NLP and CV, their development in time series domains has been constrained by data sparsity. Recent studies have revealed that large language models (LLMs) possess ro"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2310.01728","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-10-03T01:31:25Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"10c17c7b20a71707ea2554d79be47a18420a4edb84fc4f49444ea031e35f4983","abstract_canon_sha256":"d9d5c97e04950a3e8f51d224e0d54444d358a32b0977dc0d00a7860b4c28bdd8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:47.369219Z","signature_b64":"QdLAXgUEkiIZ3ueLzLkvgrnVrfsJaqqqXJ8ztn1CjeO+SpKTVdUz4JeqrMRvuKjXfl1Mgh1YfCQEPr8mHIMpDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1b288acd350970e955193e4fce6f0f37bb785b217462bafe1c85711d592a442a","last_reissued_at":"2026-05-17T23:38:47.368710Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:47.368710Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Time-LLM: Time Series Forecasting by Reprogramming Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Reprogramming time series inputs with text prototypes lets frozen large language models generate accurate forecasts without retraining.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"James Y. Zhang, Lintao Ma, Ming Jin, Pin-Yu Chen, Qingsong Wen, Shirui Pan, Shiyu Wang, Xiaoming Shi, Yuan-Fang Li, Yuxuan Liang, Zhixuan Chu","submitted_at":"2023-10-03T01:31:25Z","abstract_excerpt":"Time series forecasting holds significant importance in many real-world dynamic systems and has been extensively studied. Unlike natural language process (NLP) and computer vision (CV), where a single large model can tackle multiple tasks, models for time series forecasting are often specialized, necessitating distinct designs for different tasks and applications. While pre-trained foundation models have made impressive strides in NLP and CV, their development in time series domains has been constrained by data sparsity. Recent studies have revealed that large language models (LLMs) possess ro"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our comprehensive evaluations demonstrate that Time-LLM is a powerful time series learner that outperforms state-of-the-art, specialized forecasting models. Moreover, Time-LLM excels in both few-shot and zero-shot learning scenarios.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That reprogramming time series inputs with text prototypes and Prompt-as-Prefix successfully aligns the modalities so the frozen LLM's reasoning transfers without substantial loss of temporal structure or introduction of artifacts.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Time-LLM reprograms frozen LLMs for time series forecasting via text prototypes and Prompt-as-Prefix, outperforming specialized models in standard, few-shot, and zero-shot settings.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Reprogramming time series inputs with text prototypes lets frozen large language models generate accurate forecasts without retraining.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"0532919763307cecf322400140485a09afd9fed3ac36ad2279ab6b26a31e25fa"},"source":{"id":"2310.01728","kind":"arxiv","version":2},"verdict":{"id":"4ad2d1aa-4248-4d02-88eb-333c0038a415","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T15:59:06.403511Z","strongest_claim":"Our comprehensive evaluations demonstrate that Time-LLM is a powerful time series learner that outperforms state-of-the-art, specialized forecasting models. Moreover, Time-LLM excels in both few-shot and zero-shot learning scenarios.","one_line_summary":"Time-LLM reprograms frozen LLMs for time series forecasting via text prototypes and Prompt-as-Prefix, outperforming specialized models in standard, few-shot, and zero-shot settings.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That reprogramming time series inputs with text prototypes and Prompt-as-Prefix successfully aligns the modalities so the frozen LLM's reasoning transfers without substantial loss of temporal structure or introduction of artifacts.","pith_extraction_headline":"Reprogramming time series inputs with text prototypes lets frozen large language models generate accurate forecasts without retraining."},"references":{"count":112,"sample":[{"doi":"","year":null,"title":"Kingma and Jimmy Ba , title =","work_id":"78eb1908-e79c-467e-995d-06510492887d","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Advances in Neural Information Processing Systems , volume=","work_id":"7514ecbf-6235-4334-8b8f-970dd4d8c0ac","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"IEEE Transactions on Neural Networks and Learning Systems , year=","work_id":"b88308ed-5c51-4781-bbc9-94ba5d75c5d5","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Advances in Neural Information Processing Systems , volume=","work_id":"2a9aeb5e-58b9-4e5d-9032-ea4c8030a8d4","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining , pages=","work_id":"c8769c5e-a976-445b-b7b5-9817fff3dbc2","ref_index":7,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":112,"snapshot_sha256":"3e11eacadb3d01f8232c88ed6d247fbf633e1d8760766b693024648daaa83754","internal_anchors":4},"formal_canon":{"evidence_count":2,"snapshot_sha256":"ada81f87374c4d0d244f42efb8d431aa014556b8a8dc845e85ab06cb294ca195"},"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":"2310.01728","created_at":"2026-05-17T23:38:47.368797+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.01728v2","created_at":"2026-05-17T23:38:47.368797+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.01728","created_at":"2026-05-17T23:38:47.368797+00:00"},{"alias_kind":"pith_short_12","alias_value":"DMUIVTJVBFYO","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_16","alias_value":"DMUIVTJVBFYOSVIZ","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_8","alias_value":"DMUIVTJV","created_at":"2026-05-18T12:33:33.725879+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":33,"internal_anchor_count":33,"sample":[{"citing_arxiv_id":"2407.13278","citing_title":"Deep Time Series Models: A Comprehensive Survey and Benchmark","ref_index":214,"is_internal_anchor":true},{"citing_arxiv_id":"2410.04047","citing_title":"TS-Reasoner: Domain-Oriented Time Series Inference Agents for Reasoning and Automated Analysis","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2502.09741","citing_title":"FoNE: Precise Single-Token Number Embeddings via Fourier Features","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2510.23090","citing_title":"MAP4TS: A Multi-Aspect Prompting Framework for Time-Series Forecasting with Large Language Models","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.21975","citing_title":"Reasoning through Verifiable Forecast Actions: Consistency-Grounded RL for Financial LLMs","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2511.20577","citing_title":"MSTN: A Lightweight and Fast Model for General TimeSeries Analysis","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2605.21295","citing_title":"TimeSRL: Generalizable Time-Series Behavioral Modeling via Semantic RL-Tuned LLMs -- A Case Study in Mental Health","ref_index":23,"is_internal_anchor":true},{"citing_arxiv_id":"2605.16361","citing_title":"TailedTS: Benchmark Dataset for Heavy-Tailed Time Series Prediction and Periodicity Quantification","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2605.17340","citing_title":"Olivia: Harmonizing Time Series Foundation Models with Power Spectral Density","ref_index":8,"is_internal_anchor":true},{"citing_arxiv_id":"2506.10630","citing_title":"Time Series Forecasting as Reasoning: A Slow-Thinking Approach with Reinforced LLMs","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2509.02967","citing_title":"AR-KAN: Autoregressive-Weight-Enhanced Kolmogorov-Arnold Network for Time Series Forecasting","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2509.05215","citing_title":"BEDTime: A Unified Benchmark for Automatically Describing Time Series","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2509.12089","citing_title":"RadarPLM: Adapting Pre-trained Language Models for Marine Radar Target Detection by Selective Fine-tuning","ref_index":20,"is_internal_anchor":true},{"citing_arxiv_id":"2511.01101","citing_title":"TSVer: A Benchmark for Fact Verification Against Time-Series Evidence","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2511.08947","citing_title":"AlphaCast: A Human Wisdom-LLM Intelligence Co-Reasoning Framework for Interactive Time Series Forecasting","ref_index":11,"is_internal_anchor":true},{"citing_arxiv_id":"2511.20577","citing_title":"MSTN: A Lightweight and Fast Model for General TimeSeries Analysis","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2602.17683","citing_title":"Probabilistic NDVI Forecasting from Sparse Satellite Time Series and Weather Covariates","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2602.08318","citing_title":"Is Flow Matching Just Trajectory Replay for Sequential Data?","ref_index":52,"is_internal_anchor":true},{"citing_arxiv_id":"2602.14200","citing_title":"TS-Haystack: A Multi-Task Retrieval Benchmark for Long-Context Time-Series Reasoning","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2603.12451","citing_title":"Overcoming the Modality Gap in Context-Aided Forecasting","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2605.14422","citing_title":"What if Tomorrow is the World Cup Final? Counterfactual Time Series Forecasting with Textual Conditions","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2604.04175","citing_title":"Uncertainty-Aware Foundation Models for Clinical Data","ref_index":51,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12375","citing_title":"Agent-Based Post-Hoc Correction of Agricultural Yield Forecasts","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2604.27351","citing_title":"Heterogeneous Scientific Foundation Model Collaboration","ref_index":73,"is_internal_anchor":true},{"citing_arxiv_id":"2604.26762","citing_title":"Exploring the Potential of Probabilistic Transformer for Time Series Modeling: A Report on the ST-PT Framework","ref_index":34,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DMUIVTJVBFYOSVIZHZH443YPG6","json":"https://pith.science/pith/DMUIVTJVBFYOSVIZHZH443YPG6.json","graph_json":"https://pith.science/api/pith-number/DMUIVTJVBFYOSVIZHZH443YPG6/graph.json","events_json":"https://pith.science/api/pith-number/DMUIVTJVBFYOSVIZHZH443YPG6/events.json","paper":"https://pith.science/paper/DMUIVTJV"},"agent_actions":{"view_html":"https://pith.science/pith/DMUIVTJVBFYOSVIZHZH443YPG6","download_json":"https://pith.science/pith/DMUIVTJVBFYOSVIZHZH443YPG6.json","view_paper":"https://pith.science/paper/DMUIVTJV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.01728&json=true","fetch_graph":"https://pith.science/api/pith-number/DMUIVTJVBFYOSVIZHZH443YPG6/graph.json","fetch_events":"https://pith.science/api/pith-number/DMUIVTJVBFYOSVIZHZH443YPG6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DMUIVTJVBFYOSVIZHZH443YPG6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DMUIVTJVBFYOSVIZHZH443YPG6/action/storage_attestation","attest_author":"https://pith.science/pith/DMUIVTJVBFYOSVIZHZH443YPG6/action/author_attestation","sign_citation":"https://pith.science/pith/DMUIVTJVBFYOSVIZHZH443YPG6/action/citation_signature","submit_replication":"https://pith.science/pith/DMUIVTJVBFYOSVIZHZH443YPG6/action/replication_record"}},"created_at":"2026-05-17T23:38:47.368797+00:00","updated_at":"2026-05-17T23:38:47.368797+00:00"}