N-BEATS-MOE: N-BEATS with a Mixture-of-Experts Layer for Heterogeneous Time Series Forecasting

2025-08-10

AI Summary

To address the limited adaptability of single-model approaches in heterogeneous time-series forecasting, this paper proposes MoE-N-BEATS: an extension of the interpretable N-BEATS framework that incorporates a Mixture-of-Experts (MoE) layer with a learnable gating network for block-level dynamic expert selection and weighting. This design preserves N-BEATS’ intrinsic trend/seasonality decomposition structure and interpretability, while enabling the gating weights to explicitly reflect each time series’ dependency on distinct expert components. Extensive experiments across 12 benchmark datasets—including highly heterogeneous scenarios—demonstrate that MoE-N-BEATS consistently and significantly outperforms the original N-BEATS and multiple state-of-the-art baselines. The results validate the effectiveness and generalizability of the proposed dynamic expert adaptation mechanism in enhancing forecasting accuracy for heterogeneous time series.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Deep learning approaches are increasingly relevant for time series forecasting tasks. Methods such as N-BEATS, which is built on stacks of multilayer perceptrons (MLPs) blocks, have achieved state-of-the-art results on benchmark datasets and competitions. N-BEATS is also more interpretable relative to other deep learning approaches, as it decomposes forecasts into different time series components, such as trend and seasonality. In this work, we present N-BEATS-MOE, an extension of N-BEATS based on a Mixture-of-Experts (MoE) layer. N-BEATS-MOE employs a dynamic block weighting strategy based on a gating network which allows the model to better adapt to the characteristics of each time series. We also hypothesize that the gating mechanism provides additional interpretability by identifying which expert is most relevant for each series. We evaluate our method across 12 benchmark datasets against several approaches, achieving consistent improvements on several datasets, especially those composed of heterogeneous time series.

Problem

• Extends N-BEATS for heterogeneous time series forecasting
• Improves adaptability via Mixture-of-Experts dynamic weighting
• Enhances interpretability by identifying relevant experts per series

Innovation

• Extends N-BEATS with Mixture-of-Experts layer
• Uses dynamic block weighting strategy
• Improves interpretability via expert gating