Pith. sign in

REVIEW 10 cited by

AutoTimes: Autoregressive Time Series Forecasters via Large Language Models

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2402.02370 v4 pith:WTNLGZXK submitted 2024-02-04 cs.LG cs.CL

Yong Liu , Guo Qin , Xiangdong Huang , Jianmin Wang , Mingsheng Long This is my paper
classification cs.LGcs.CL
keywords seriestimeautotimeslanguagellmsmodelsautoregressiveforecasters
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
abstract

Foundation models of time series have not been fully developed due to the limited availability of time series corpora and the underexploration of scalable pre-training. Based on the similar sequential formulation of time series and natural language, increasing research demonstrates the feasibility of leveraging large language models (LLM) for time series. Nevertheless, the inherent autoregressive property and decoder-only architecture of LLMs have not been fully considered, resulting in insufficient utilization of LLM abilities. To fully revitalize the general-purpose token transition and multi-step generation capability of large language models, we propose AutoTimes to repurpose LLMs as autoregressive time series forecasters, which projects time series into the embedding space of language tokens and autoregressively generates future predictions with arbitrary lengths. Compatible with any decoder-only LLMs, the consequent forecaster exhibits the flexibility of the lookback length and scalability with larger LLMs. Further, we formulate time series as prompts, extending the context for prediction beyond the lookback window, termed in-context forecasting. By introducing LLM-embedded textual timestamps, AutoTimes can utilize chronological information to align multivariate time series. Empirically, AutoTimes achieves state-of-the-art with 0.1% trainable parameters and over $5\times$ training/inference speedup compared to advanced LLM-based forecasters. Code is available at this repository: https://github.com/thuml/AutoTimes.

Discussion (0). Sign in to comment.

Forward citations

Cited by 10 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Multi-Source Dynamic Graph Learning for Compound-Flood Forecasting in Managed Coastal Systems

    cs.LG 2026-08 conditional novelty 6.0 of 10

    An anchored forecaster that adds bounded, regime-gated corrections from a dynamic multi-station graph to a local temporal forecast improves sustained high-water plateau prediction in South Florida without harming rout...

  2. A Cost-Effective Multimodal LLM Reasoning Framework for Question Answering over Irregular Clinical Time Series

    cs.AI 2026-07 conditional novelty 6.0 of 10

    ClinPRISM reaches 49.83% average accuracy on CLIR-Bench irregular clinical time-series QA using a 4B LLM, 16 temporal tokens, and 0.15 s/question.

  3. CLIR-Bench: Benchmarking Multimodal Question Answering over Irregular Clinical Time Series

    cs.CL 2026-07 conditional novelty 6.0 of 10

    CLIR-Bench shows generalist and time-series LLMs struggle to ground clinical answers in sparse irregular ICU evidence, with top accuracy near 50% and weak causal evidence use.

  4. LLM-Guided Task-Semantic Field Factorization for Industrial Process Forecasting

    cs.LG 2026-07 conditional novelty 6.0 of 10

    TSF converts process variable documents into frozen semantic vectors that scale the numerical input window before a time-series backbone, yielding average MAE reductions of 2.9–3.6% across industrial forecasting tasks.

  5. LLM-Guided Measurement Credibility Correction for Trustworthy Industrial Process Inference

    eess.SY 2026-07 conditional novelty 6.0 of 10

    LLM-derived measurement semantics enable a lightweight pre-inference correction step that reduces industrial prediction MAE by 30.7% on real tests and 80.3% under controlled sensor corruption.

  6. From Bias to Behavior: Learning Bull-Bear Market Dynamics with Contrastive Modeling

    cs.LG 2025-07 reject novelty 6.0 of 10

    A contrastive model, B4, jointly learns price and news representations split into bullish and bearish camps, claiming better trend prediction and interpretable bias dynamics.

  7. Output Scaling: YingLong-Delayed Chain of Thought in a Large Pretrained Time Series Forecasting Model

    cs.LG 2025-05 conditional novelty 6.0 of 10

    Forecasting with a non-causal encoder-only model improves fixed-horizon accuracy when the model is asked to output extra future tokens, an effect the authors call delayed chain-of-thought.

  8. Teaching Time Series to See and Speak: Forecasting with Aligned Visual and Textual Perspectives

    cs.LG 2025-06 reject novelty 5.0 of 10

    TimesCLIP aligns image-based and text-based views of the same time series via contrastive learning to improve forecasting accuracy on several benchmarks, but the full multimodal model is not used on two of the six lon...

  9. PIPE: Physics-Informed Position Encoding for Alignment of Satellite Images and Time Series

    cs.LG 2025-05 conditional novelty 5.0 of 10

    A physics-informed positional encoding that embeds timestamps and latitude/longitude into a vision language model improves typhoon intensity and track forecasting on the Digital Typhoon dataset.

  10. Analyzing Patient Daily Movement Behavior Dynamics Using Two-Stage Encoding Model

    cs.AI 2025-02 reject novelty 4.0 of 10

    A two-stage pipeline encodes daily home-activity text with MiniLM, clusters the embeddings, and applies PageRank to derive per-patient behavior vectors.

Pith tools