PatchFormer: A Patch-Based Time Series Foundation Model with Hierarchical Masked Reconstruction and Cross-Domain Transfer Learning for Zero-Shot Multi-Horizon Forecasting
cs.LG, eess.SP
Submitted: 2026-01-28
Updated: 2026-09-06
Comments: Withdrawn by mutual agreement of all authors due to critical data leakage identified in the temporal train-test splitting and masked reconstruction pipeline, which invalidates the reported zero-shot forecasting benchmarks and core conclusions
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- TimeGPT-1
- Chronos: Learning the Language of Time Series
- MOMENT: A Family of Open Time-series Foundation Models
- Lag-Llama: Towards Foundation Models for Probabilistic Time Series Forecasting
- A decoder-only foundation model for time-series forecasting
- An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling
- WaveNet: A Generative Model for Raw Audio
- Scaleformer: Iterative Multi-scale Refining Transformers for Time Series Forecasting
- A Time Series is Worth 64 Words: Long-term Forecasting with Transformers
- Scaling Laws for Neural Language Models
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks