Anti-Aliasing Matters: A Dynamic Network for Time Series Forecasting

Real-world time series inherently exhibit complex temporal patterns. Within chaotic systems, significant mixing and entanglement occur between different time-varying modes. Given that time series exhibit distinctly different patterns at various sampling scales, downsampling to extract multiscale features is a common approach. However, conventional downsampling causes high-frequency components in the original signal, those exceeding the new Nyquist frequency, to undergo spectral folding. This erroneously introduces spurious low-frequency patterns, perceived as low-frequency noise, thereby leading to the aliasing problem. To address this problem, we propose a Decomposition-Prevention-Fusion architecture framework called DMANet, which introduces the Dynamic Multiscale Anti-Aliasing Network. Specifically, DMANet comprises two key components: Multiscale Convolutional Downsampling, designed to capture temporal dependencies and inter-channel interactions, and an Anti-Aliasing Operation, which includes Pre-Sampling Anti-Aliasing Filtering and Post-Sampling Interpolation. These designs guarantee the fidelity of multiscale features before and after downsampling. We show that by mitigating the risk of aliasing, our proposed simple convolutional downsampling architecture achieves performance competitive with common baselines. Our codes are available at https://github.com/HenryZhou2024/DMANet.

Time series data, such as weather records, electricity usage, and industrial sensor signals, often contain complex patterns that change across different time scales. Modern AI models commonly compress these signals to capture both short-term and long-term trends. However, this process can unintentionally distort the original data and introduce misleading patterns, causing forecasting models to learn incorrect information. In this work, we show how these distortions affect time series forecasting and propose DMANet, a simple framework designed to reduce them. Our method preserves important temporal patterns during multiscale processing through adaptive filtering and feature reconstruction. Experiments on real-world datasets show that reducing signal distortion leads to more accurate and reliable forecasting performance.