LAST SToP For Modeling Asynchronous Time Series

TL;DR

This paper introduces a novel prompt design and stochastic soft prompting mechanism for Large Language Models to effectively analyze asynchronous time series data, enhancing tasks like forecasting, anomaly detection, and data imputation.

Contribution

The paper presents a new prompt design and a stochastic soft prompting method that significantly improve LLM performance on asynchronous time series tasks.

Findings

Achieves state-of-the-art results on real-world datasets.

Outperforms existing fine-tuning methods like QLoRA.

Enhances LLM capabilities in anomaly detection and data imputation.

Abstract

We present a novel prompt design for Large Language Models (LLMs) tailored to Asynchronous Time Series. Unlike regular time series, which assume values at evenly spaced time points, asynchronous time series consist of timestamped events occurring at irregular intervals, each described in natural language. Our approach effectively utilizes the rich natural language of event descriptions, allowing LLMs to benefit from their broad world knowledge for reasoning across different domains and tasks. This allows us to extend the scope of asynchronous time series analysis beyond forecasting to include tasks like anomaly detection and data imputation. We further introduce Stochastic Soft Prompting, a novel prompt-tuning mechanism that significantly improves model performance, outperforming existing fine-tuning methods such as QLoRA. Through extensive experiments on real world datasets, we demonstrate that our approach achieves state-of-the-art performance across different tasks and datasets.