aiqbench analysis
NPMixer addresses the complexity of multivariate time series forecasting by introducing a hierarchical architecture that adaptively decomposes signals using learnable wavelet transforms and MLP-based mixing blocks. The model effectively captures both local temporal dynamics within patches and global dependencies across variables, demonstrating superior performance on seven benchmark datasets compared to existing state-of-the-art methods. This research advances the field of sequence modeling through novel signal decomposition techniques tailored for data-dependent filter coefficient learning.
NPMixer: Hierarchical Neighboring Patch Mixing for Time Series Forecasting
arXiv:2605.07476v2 Announce Type: replace Abstract: Multivariate time series forecasting remains a challenge due to the complexity of local temporal dynamics and global dependencies across multiple variables. In this paper, we propose \textbf{N}eighboring \textbf{P}atching \textbf{Mixer} (\textbf{NPMixer}), a hierarchical architecture featuring a Learnable Stationary Wavelet Transform that adaptively learns filter coefficients to decompose signals into trend and detail components in a data-dependent manner. Our framework introduces a Neighboring Mixer Block that captures local temporal dynamics through a series of hierarchical MLP layers operating on non-overlapping patches. Specifically, the mixer block utilizes MLPs to learn temporal patterns within and across these patches, expanding the receptive field to capture multi-scale dependencies. A Channel-Mixing Encoder is applied to high-frequency components to learn channel correlations while preserving the stability of the underlying global trend. Extensive experiments on seven benchmark datasets demonstrate that NPMixer consistently outperforms state-of-the-art models, achieving better performance in 20 out of 28 ($71.4\%$) evaluated experimental setups for MSE.
Abstract:Multivariate time series forecasting remains a challenge due to the complexity of local temporal dynamics and global dependencies across multiple variables. In this paper, we propose \textbf{N}eighboring \textbf{P}atching \textbf{Mixer} (\textbf{NPMixer}), a hierarchical architecture featuring a Learnable Stationary Wavelet Transform that adaptively learns filter coefficients to decompose signals into trend and detail components in a data-dependent manner. Our framework introduces a Neighboring Mixer Block that captures local temporal dynamics through a series of hierarchical MLP layers operating on non-overlapping patches. Specifically, the mixer block utilizes MLPs to learn temporal patterns within and across these patches, expanding the receptive field to capture multi-scale dependencies. A Channel-Mixing Encoder is applied to high-frequency components to learn channel correlations while preserving the stability of the underlying global trend. Extensive experiments on seven benchmark datasets demonstrate that NPMixer consistently outperforms state-of-the-art models, achieving better performance in 20 out of 28 ($71.4\%$) evaluated experimental setups for MSE.
Machine Learning (cs.LG)
arXiv:2605.07476 [cs.LG]
(or arXiv:2605.07476v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2605.07476 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 242 words
Related coverage
A Foundation Model for Instruction-Conditioned In-Context Time Series Tasks
arXiv · Sep 24, 2026
Introducing Timestral
Mistral AI · Oct 15, 2026
Real-Time Intelligence with IBM Time Series Models on Confluent
Hugging Face · Sep 2, 2026
PatchFormer: A Patch-Based Time Series Foundation Model with Hierarchical Masked Reconstruction and Cross-Domain Transfer Learning for Zero-Shot Multi-Horizon Forecasting
arXiv · Sep 9, 2026