MR-ImagenTime: Multi-Resolution Time Series Generation through Dual Image Representations

TL;DR

MR-ImagenTime introduces a multi-resolution framework for time series generation that effectively handles variable-length inputs and multi-scale patterns, outperforming existing models.

Contribution

It presents MR-CDM, a novel framework combining hierarchical trend decomposition, adaptive embeddings, and multi-scale diffusion for improved time series forecasting.

Findings

Significantly outperforms state-of-the-art models like CSDI and Informer.

Reduces MAE and RMSE by approximately 6-10%.

Effective on four real-world datasets.

Abstract

Time series forecasting is vital across many domains, yet existing models struggle with fixed-length inputs and inadequate multi-scale modeling. We propose MR-CDM, a framework combining hierarchical multi-resolution trend decomposition, an adaptive embedding mechanism for variable-length inputs, and a multi-scale conditional diffusion process. Evaluations on four real-world datasets demonstrate that MR-CDM significantly outperforms state-of-the-art baselines (e.g., CSDI, Informer), reducing MAE and RMSE by approximately 6-10 to a certain degree.