Learning to Difference: Adaptive Reversible Differencing (AdaRDiff) for Time Series Forecasting

• URL: http://arxiv.org/abs/2608.28134v1
• Date: Fri, 28 Aug 2026 09:51:08 GMT
• Title: Learning to Difference: Adaptive Reversible Differencing (AdaRDiff) for Time Series Forecasting
• Abstract summary: Trends and seasonality introduce complex temporal structure that challenges learning-based forecasting models.<n>We propose textbfunderlineAdaptive textbfunderlineReversible textbfunderlineDifferencing textbf(AdaRDiff).<n>AdaRDiff attains state-of-the-art forecast accuracy across eight benchmarks spanning electricity, weather, traffic, and energy, at negligible parameter cost.
• Score: 2.370553892492642
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Reliable long-horizon time series forecasting is an important yet difficult problem. Trends and seasonality introduce complex temporal structure that challenges learning-based forecasting models. Differencing, which subtracts nearby past values to remove such structure, is the classical remedy, but its reliance on hand-picked orders and periods has kept it largely absent from recent deep architectures. We propose \textbf{\underline{Ada}}ptive \textbf{\underline{R}}eversible \textbf{\underline{Diff}}erencing \textbf{(AdaRDiff)}, a generalized differencing approach that uses learnable weights to simplify the series through weighted differencing with previous time instants. This yields stabilized residuals on which forecasting is performed, after which the removed components are restored autoregressively to reconstruct the forecast, capturing trend and seasonality jointly through a single operator. This reconstruction admits a closed-form convolutional expression, which parallelizes on GPU and yields up to $33.7\times$ speedup over the naive recurrence. We furthermore rely on a two-phase training schedule that separates temporal structure discovery from reconstruction learning, as suggested by a theoretical analysis of the gradient when using a linear forecasting model. AdaRDiff attains state-of-the-art forecast accuracy across eight benchmarks spanning electricity, weather, traffic, and energy, at negligible parameter cost. Furthermore, it is designed as a plug-and-play module: integrating AdaRDiff improves eight diverse backbones, from linear models to Transformers, in the large majority of cases, by up to $25.9\%$ with a linear backbone and $18.3\%$ with iTransformer.

Related papers

• GatedLinear: Adaptive Routing of Complementary Linear Bases for Time Series Forecasting [11.155508081323694]
  Time series forecasting requires models to capture diverse, often mutually exclusive, temporal dynamics.<n>Recent deep learning models have improved accuracy, but they typically force these diverse patterns through a single computational backbone.<n>We propose GatedLinear: a lightweight framework that frames forecasting as the adaptive routing of complementary linear bases.
  arXiv  Detail & Related papers  (2026-07-10T15:42:01Z)
• StateFlow: Dual-State Recurrent Modeling for Long-Horizon Time Series Forecasting [21.257843115741256]
  StateFlow is a recurrent forecasting framework that uses VARNN as a dual-state recurrent backbone to capture two complementary signals from the lookback sequence.<n>A chunk-based decoder separately summarizes these trajectories and maps them to the future horizon for direct multi-step forecasting.
  arXiv  Detail & Related papers  (2026-06-30T21:26:09Z)
• One Step Closer to Ground Truth: A Multi-Scale Residual-Aware Representation Learning Pipeline for Predicting Time Series Data [6.704099492625706]
  We introduce a two-stage, model-agnostic framework that explicitly decouples forecasting and residual learning into distinct stages of representation learning.<n>By formalizing this pipeline as a hypothesis space expansion, our framework addresses approximation limitations inherent in single-stage architectures.
  arXiv  Detail & Related papers  (2026-06-09T10:31:54Z)
• Factorize to Generalize: Retrieval-Guided Invariant-Dynamic Decomposition for Time Series Forecasting [64.2640656619583]
  Time series foundation models (TSFMs) have recently achieved strong zero-shot forecasting performance through large-scale pretraining and retrieval-augmented prediction.<n>We propose a Retrieval-guided Invariant-Dynamic DEcomposition framework for time series forecasting.
  arXiv  Detail & Related papers  (2026-05-24T07:31:55Z)
• Next-Acceleration-Scale Prediction for Autoregressive MRI Reconstruction [52.32112533846212]
  MRI reconstruction is an inherently ill-posed inverse problem, since incomplete measurements admit many plausible solutions.<n>We address this limitation by moving reconstruction to discrete multi-scale latent space and posing it as autoregressive next-acceleration-scale prediction.<n>We show that our approach delivers improved reconstruction performance across diverse sampling patterns under extreme undersampling.
  arXiv  Detail & Related papers  (2026-05-19T04:40:50Z)
• Three-Stage Learning Unlocks Strong Performance in Simple Models for Long-Term Time Series Forecasting [6.603558172147427]
  STAIR aims to unlock the capacity of simple temporal mapping models without introducing complex architectural modules.<n>It first learns common temporal dynamics across variables through a shared temporal mapping, then adapts the shared model to each variable via channel-wise fine-tuning to capture variable-specific patterns, and finally complements the backbone with cross-variable information through residual learning.
  arXiv  Detail & Related papers  (2026-05-13T15:36:46Z)
• PnP-Corrector: A Universal Correction Framework for Coupled Spatiotemporal Forecasting [62.22445401483844]
  Existing methods are severely constrained by the bottleneck of compounding errors.<n>We propose a universal framework called -Cast-Corrector (Plug-and-Play Corrector)<n>Our method significantly enhances the long-term stability of coupled forecasting systems.
  arXiv  Detail & Related papers  (2026-05-09T13:12:33Z)
• SSR: A Training-Free Approach for Streaming 3D Reconstruction [53.19807901585702]
  Self-expressive Sequence Regularization (SSR) is a plug-and-play operator that enforces Grassmannian sequence regularity during inference.<n>We show that SSR consistently reduces drift and improves reconstruction quality across multiple streaming 3D reconstruction tasks.
  arXiv  Detail & Related papers  (2026-03-16T02:55:14Z)
• Forecasting as Rendering: A 2D Gaussian Splatting Framework for Time Series Forecasting [79.37674445572462]
  Time series forecasting (TSF) remains a challenging problem due to the intricate entanglement of intraperiod-fluctuations and interperiod-trends.<n>Treating reshaped tensors as static images results in a topological mismatch.<n> relying on uniform fixed-size representations allocates modeling capacity inefficiently.<n>TimeGS is a novel framework that fundamentally shifts the forecasting paradigm from regression to 2D generative rendering.
  arXiv  Detail & Related papers  (2026-02-10T14:13:36Z)
• Timer-XL: Long-Context Transformers for Unified Time Series Forecasting [67.83502953961505]
  We present Timer-XL, a causal Transformer for unified time series forecasting.<n>Based on large-scale pre-training, Timer-XL achieves state-of-the-art zero-shot performance.
  arXiv  Detail & Related papers  (2024-10-07T07:27:39Z)
• Parsimony or Capability? Decomposition Delivers Both in Long-term Time Series Forecasting [46.63798583414426]
  Long-term time series forecasting (LTSF) represents a critical frontier in time series analysis. Our study demonstrates, through both analytical and empirical evidence, that decomposition is key to containing excessive model inflation. Remarkably, by tailoring decomposition to the intrinsic dynamics of time series data, our proposed model outperforms existing benchmarks.
  arXiv  Detail & Related papers  (2024-01-22T13:15:40Z)
• PatchMixer: A Patch-Mixing Architecture for Long-Term Time Series Forecasting [3.7244649185886978]
  We propose PatchMixer, a novel CNN-based model. It introduces a permutation-variant convolutional structure to preserve temporal information. Compared with the state-of-the-art method and the best-performing CNN, PatchMixer yields $3.9%$ and $21.2%$ relative improvements.
  arXiv  Detail & Related papers  (2023-10-01T12:47:59Z)
• Koopa: Learning Non-stationary Time Series Dynamics with Koopman Predictors [85.22004745984253]
  Real-world time series are characterized by intrinsic non-stationarity that poses a principal challenge for deep forecasting models. We tackle non-stationary time series with modern Koopman theory that fundamentally considers the underlying time-variant dynamics. We propose Koopa as a novel Koopman forecaster composed of stackable blocks that learn hierarchical dynamics.
  arXiv  Detail & Related papers  (2023-05-30T07:40:27Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.