REVIEW 12 cited by
Large Language Models Are Zero-Shot Time Series Forecasters
Not yet reviewed by Pith; the record is open.
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
read the original abstract
By encoding time series as a string of numerical digits, we can frame time series forecasting as next-token prediction in text. Developing this approach, we find that large language models (LLMs) such as GPT-3 and LLaMA-2 can surprisingly zero-shot extrapolate time series at a level comparable to or exceeding the performance of purpose-built time series models trained on the downstream tasks. To facilitate this performance, we propose procedures for effectively tokenizing time series data and converting discrete distributions over tokens into highly flexible densities over continuous values. We argue the success of LLMs for time series stems from their ability to naturally represent multimodal distributions, in conjunction with biases for simplicity, and repetition, which align with the salient features in many time series, such as repeated seasonal trends. We also show how LLMs can naturally handle missing data without imputation through non-numerical text, accommodate textual side information, and answer questions to help explain predictions. While we find that increasing model size generally improves performance on time series, we show GPT-4 can perform worse than GPT-3 because of how it tokenizes numbers, and poor uncertainty calibration, which is likely the result of alignment interventions such as RLHF.
Forward citations
Cited by 12 Pith papers
-
LLM-Guided Task-Semantic Field Factorization for Industrial Process Forecasting
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.
-
Not Just What, But When: Integrating Irregular Intervals to LLM for Sequential Recommendation
IntervalLLM integrates irregular time intervals into an LLM recommender via interval embeddings and interval-infused attention, improving next-item Hit Rate@1 on three benchmarks and adding a new interval-perspective ...
-
Causal Graph Fuzzy LLMs: A First Introduction and Applications in Time Series Forecasting
CGF-LLM combines fuzzy time series and PCMCI causal graphs into text input for fine-tuned GPT-2, reporting improved one-step-ahead forecast NRMSE and a reduction in token count on four datasets.
-
Towards Interpretable Time Series Foundation Models
After fine-tuning on 180 synthetic mean-reverting series annotated by a large multimodal model, small Qwen models can describe trend direction, noise intensity, and extremum location in natural language.
-
Evaluation of a Foundational Model and Stochastic Models for Forecasting Sporadic or Spiky Production Outages of High-Performance Machine Learning Services
On seven years of monthly production outage counts from a large ML service, a fine-tuned TimesFM foundation model beats moving-average and autoregressive baselines for total outages, but per root cause the best model varies.
-
Random Initialization Can't Catch Up: The Advantage of Language Model Transfer for Time Series Forecasting
Pretrained T5 language weights give a persistent validation-loss advantage over random initialization for low-data time series forecasting, and the advantage does not vanish within the training budget.
-
When can isotropy help adapt LLMs' next word prediction to numerical domains?
Using a log-linear model and Jacobian analysis, the paper claims isotropy in LLM hidden embeddings stabilizes the softmax partition function and improves time-series forecasting, with illustrative experiments on five ...
-
On the Feasibility of Vision-Language Models for Time-Series Classification
Fine-tuning a vision-language model for one or two epochs on line plots plus text can classify several UCR time-series datasets, but the 'competitive' claim is never tested against standard baselines.
-
Technology-assisted Personalized Yoga for Better Health -- Challenges and Outlook
A vision statement for personalized yoga decision support whose claimed content cannot be verified, because the supplied full text is a different paper on maritime drift prediction.
-
Towards Foundation Auto-Encoders for Time-Series Anomaly Detection
A univariate VAE with dilated convolutions is proposed as a simple 'foundation' model for time-series anomaly detection, with preliminary zero-shot experiments on two datasets.
-
FD-LLM: Large Language Model for Fault Diagnosis of Machines
Open-source LLMs fine-tuned with LoRA on FFT spectra encoded as text achieve 99%+ accuracy on CWRU bearing fault classification and generally beat a 1D-CNN under cross-load transfer, while cross-component transfer rem...
-
Fine-Tuning Pre-trained Large Time Series Models for Prediction of Wind Turbine SCADA Data
Fine-tuning the pre-trained Timer model on a single wind turbine's SCADA data yields the best prediction accuracy across the whole wind farm, while the model does not always beat simpler baselines on larger datasets.
Discussion (0). Continue with ORCID to comment.