🤖 AI Summary
To address noise sensitivity, temporal information loss, and insufficient inter-variable dependency modeling in long-term time series forecasting (LTSF), this paper proposes PatchMLP—a lightweight, MLP-based architecture. Departing from complex Transformer designs, it theoretically and empirically establishes patch-based decomposition as the key driver of LTSF performance gains. PatchMLP introduces a dual-path noise modeling paradigm: channel-mixing and channel-independent branches that explicitly capture both smooth trends and residual noise. Furthermore, it incorporates cross-variable semantic interaction modules to enhance multi-variate collaborative representation learning. Built upon patch decomposition, sliding-average smoothing, and fully connected layers, PatchMLP achieves significantly lower parameter count and computational cost than state-of-the-art Transformer-based methods. Extensive experiments on multiple real-world benchmarks demonstrate its superior forecasting accuracy, establishing new state-of-the-art performance in LTSF.
📝 Abstract
Recent studies have attempted to refine the Transformer architecture to demonstrate its effectiveness in Long-Term Time Series Forecasting (LTSF) tasks. Despite surpassing many linear forecasting models with ever-improving performance, we remain skeptical of Transformers as a solution for LTSF. We attribute the effectiveness of these models largely to the adopted Patch mechanism, which enhances sequence locality to an extent yet fails to fully address the loss of temporal information inherent to the permutation-invariant self-attention mechanism. Further investigation suggests that simple linear layers augmented with the Patch mechanism may outperform complex Transformer-based LTSF models. Moreover, diverging from models that use channel independence, our research underscores the importance of cross-variable interactions in enhancing the performance of multivariate time series forecasting. The interaction information between variables is highly valuable but has been misapplied in past studies, leading to suboptimal cross-variable models. Based on these insights, we propose a novel and simple Patch-based MLP (PatchMLP) for LTSF tasks. Specifically, we employ simple moving averages to extract smooth components and noise-containing residuals from time series data, engaging in semantic information interchange through channel mixing and specializing in random noise with channel independence processing. The PatchMLP model consistently achieves state-of-the-art results on several real-world datasets. We hope this surprising finding will spur new research directions in the LTSF field and pave the way for more efficient and concise solutions.