MemCast: Memory-Driven Time Series Forecasting with Experience-Conditioned Reasoning

TL;DR

MemCast introduces a memory-augmented framework for time series forecasting that leverages experience organization and reasoning to improve prediction accuracy and enable continual learning.

Contribution

It presents a novel experience-conditioned reasoning approach with hierarchical memory and dynamic confidence adaptation for improved TSF performance.

Findings

MemCast outperforms previous methods on multiple datasets.

The hierarchical memory improves experience utilization.

Dynamic confidence updates enable continual evolution.

Abstract

Time series forecasting (TSF) plays a critical role in decision-making for many real-world applications. Recently, LLM-based forecasters have made promising advancements. Despite their effectiveness, existing methods often lack explicit experience accumulation and continual evolution. In this work, we propose MemCast, a learning-to-memory framework that reformulates TSF as an experience-conditioned reasoning task. Specifically, we learn experience from the training set and organize it into a hierarchical memory. This is achieved by summarizing prediction results into historical patterns, distilling inference trajectories into reasoning wisdom, and inducing extracted temporal features into general laws. Furthermore, during inference, we leverage historical patterns to guide the reasoning process and utilize reasoning wisdom to select better trajectories, while general laws serve as criteria for reflective iteration. Additionally, to enable continual evolution, we design a dynamic confidence adaptation strategy that updates the confidence of individual entries without leaking the test set distribution. Extensive experiments on multiple datasets demonstrate that MemCast consistently outperforms previous methods, validating the effectiveness of our approach. Our code is available at https://github.com/Xiaoyu-Tao/MemCast-TS.