mlsanityData to 14 Oct 2026

Feature-aligned N-BEATS with Sinkhorn divergence

arXiv 2305.15196 ↗First posted 24 May 2023cs.LG cs.AI math.OCcited by 0

Not on the map. Why?

Abstract (arXiv)

We propose Feature-aligned N-BEATS as a domain-generalized time series forecasting model. It is a nontrivial extension of N-BEATS with doubly residual stacking principle (Oreshkin et al. [45]) into a representation learning framework. In particular, it revolves around marginal feature probability measures induced by the intricate composition of residual and feature extracting operators of N-BEATS in each stack and aligns them stack-wise via an approximate of an optimal transport distance referred to as the Sinkhorn divergence. The training loss consists of an empirical risk minimization from multiple source domains, i.e., forecasting loss, and an alignment loss calculated with the Sinkhorn divergence, which allows the model to learn invariant features stack-wise across multiple source data sequences while retaining N-BEATS's interpretable design and forecasting power. Comprehensive experimental evaluations with ablation studies are provided and the corresponding results demonstrate the proposed model's forecasting and generalization capabilities.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 42 of its 77 references that are in the map. First posted on arXiv 24 May 2023.

Builds on

42 of 77 references.

References · 42

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2022Generative Adversarial Networkscited by 5,627
  3. 2015Domain-Adversarial Training of Neural Networkscited by 3,044
  4. 2018UMAP: Uniform Manifold Approximation and Projection for Dimension Reductioncited by 2,898
  5. 2020Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecastingcited by 2,169
  6. 2014Unsupervised Domain Adaptation by Backpropagationcited by 1,917
  7. 2021Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecastingcited by 1,676
  8. 2018Spectral Normalization for Generative Adversarial Networkscited by 1,636
  9. 2022Are Transformers Effective for Time Series Forecasting?cited by 1,424
  10. 2022FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecastingcited by 1,210
  11. 2017DeepAR: Probabilistic Forecasting with Autoregressive Recurrent Networkscited by 1,013
  12. 2021Generalizing to Unseen Domains: A Survey on Domain Generalizationcited by 766
30 more
  1. 2019Temporal Fusion Transformers for Interpretable Multi-horizon Time Series Forecastingcited by 761
  2. 2019N-BEATS: Neural basis expansion analysis for interpretable time series forecastingcited by 753
  3. 2018Deep Visual Domain Adaptation: A Surveycited by 635
  4. 2013Domain Generalization via Invariant Feature Representationcited by 623
  5. 2018Lipschitz regularity of deep neural networks: analysis and efficient estimationcited by 415
  6. 2021SCINet: Time Series Modeling and Forecasting with Sample Convolution and Interactioncited by 338
  7. 2017Unified Deep Supervised Domain Adaptation and Generalizationcited by 323
  8. 2017Learning Generative Models with Sinkhorn Divergencescited by 314
  9. 2018BRITS: Bidirectional Recurrent Imputation for Time Seriescited by 290
  10. 2019Sliced Wasserstein Discrepancy for Unsupervised Domain Adaptationcited by 235
  11. 2016Stochastic Optimization for Large-scale Optimal Transportcited by 189
  12. 2019Recurrent Neural Networks for Time Series Forecasting: Current Status and Future Directionscited by 172
  13. 2019Domain Generalization Using a Mixture of Multiple Latent Domainscited by 170
  14. 2015On Wasserstein Two Sample Testing and Related Families of Nonparametric Testscited by 166
  15. 2022ETSformer: Exponential Smoothing Transformers for Time-series Forecastingcited by 148
  16. 2022N-HiTS: Neural Hierarchical Interpolation for Time Series Forecastingcited by 146
  17. 2019Generalizing to unseen domains via distribution matchingcited by 110
  18. 2015Scatter Component Analysis: A Unified Framework for Domain Adaptation and Domain Generalizationcited by 96
  19. 2020Multi-Source Deep Domain Adaptation with Weak Supervision for Time-Series Sensor Datacited by 82
  20. 2020Faster Wasserstein Distance Estimation with the Sinkhorn Divergencecited by 73
  21. 2020COT-GAN: Generating Sequential Data via Causal Optimal Transportcited by 71
  22. 2020Meta-learning framework with applications to zero-shot time-series forecastingcited by 70
  23. 2021Domain Adaptation for Time Series Forecasting via Attention Sharingcited by 68
  24. 2019Probabilistic Forecasting with Temporal Convolutional Neural Networkcited by 63
  25. 2020Domain Generalization via Optimal Transport with Metric Similarity Learningcited by 27
  26. 2021On the benefits of representation regularization in invariance based domain generalizationcited by 17
  27. 2021Yformer: U-Net Inspired Transformer Architecture for Far Horizon Time Series Forecastingcited by 11
  28. 2020Optimal Transport losses and Sinkhorn algorithm with general convex regularizationcited by 10
  29. 2021Reliable Estimation of KL Divergence using a Discriminator in Reproducing Kernel Hilbert Spacecited by 7
  30. 2022Time-Series Domain Adaptation via Sparse Associative Structure Alignment: Learning Invariance and Variancecited by 4

Built on by

Not cited yet.