SPADE-S: A Sparsity-Robust Foundational Forecaster

2025-07-24

AI Summary

To address systematic prediction bias in low-magnitude and sparse time series—arising from magnitude heterogeneity and data sparsity—this paper proposes a robust time-series forecasting architecture. Methodologically, it introduces: (1) a debiased loss function to mitigate gradient imbalance induced by amplitude disparities; (2) a dynamic sparsity-aware sampling strategy to enhance training stability on sparse sequences; and (3) a temporal-adaptive encoder to strengthen multi-scale pattern representation. This is the first work to systematically tackle performance degradation of deep forecasting models on heterogeneous time series. Evaluated on three large-scale e-commerce datasets containing 3 million to 700 million series, the method achieves P90 accuracy gains of 2.21%–6.58%, a P50 improvement of 1.95%, and up to 15% single-point accuracy gain over state-of-the-art baselines.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Search and Optimization: Learning to Search

Application Category

• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Despite significant advancements in time series forecasting, accurate modeling of time series with strong heterogeneity in magnitude and/or sparsity patterns remains challenging for state-of-the-art deep learning architectures. We identify several factors that lead existing models to systematically underperform on low-magnitude and sparse time series, including loss functions with implicit biases toward high-magnitude series, training-time sampling methods, and limitations of time series encoding methods. SPADE-S is a robust forecasting architecture that significantly reduces magnitude- and sparsity-based systematic biases and improves overall prediction accuracy. Empirical results demonstrate that SPADE-S outperforms existing state-of-the-art approaches across a diverse set of use cases in demand forecasting. In particular, we show that, depending on the quantile forecast and magnitude of the series, SPADE-S can improve forecast accuracy by up to 15%. This results in P90 overall forecast accuracy gains of 2.21%, 6.58%, and 4.28%, and P50 forecast accuracy gains of 0.92%, 0.77%, and 1.95%, respectively, for each of three distinct datasets, ranging from 3 million to 700 million series, from a large online retailer.

Problem

• Addresses biases in forecasting low-magnitude sparse time series
• Improves accuracy for heterogeneous magnitude and sparsity patterns
• Enhances demand forecasting across diverse large-scale datasets

Innovation

• Robust architecture for heterogeneous time series
• Reduces magnitude and sparsity biases
• Improves forecast accuracy up to 15%