Dynamic Tokenization via Reinforcement Patching: End-to-end Training and Zero-shot Transfer

• URL: http://arxiv.org/abs/2603.26097v1
• Date: Fri, 27 Mar 2026 06:08:30 GMT
• Title: Dynamic Tokenization via Reinforcement Patching: End-to-end Training and Zero-shot Transfer
• Abstract summary: We propose ReinPatch, a framework to jointly optimize a sequence patching policy and its downstream sequence backbone model.<n>By formulating patch boundary placement as a discrete decision process optimized via Group Relative Policy Gradient (GRPG), ReinPatch bypasses the need for continuous relaxations.<n>We evaluate ReinPatch on time-series forecasting datasets, where it demonstrates compelling performance compared to state-of-the-art data-driven patching strategies.
• Score: 19.88776013507579
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Efficiently aggregating spatial or temporal horizons to acquire compact representations has become a unifying principle in modern deep learning models, yet learning data-adaptive representations for long-horizon sequence data, especially continuous sequences like time series, remains an open challenge. While fixed-size patching has improved scalability and performance, discovering variable-sized, data-driven patches end-to-end often forces models to rely on soft discretization, specific backbones, or heuristic rules. In this work, we propose Reinforcement Patching (ReinPatch), the first framework to jointly optimize a sequence patching policy and its downstream sequence backbone model using reinforcement learning. By formulating patch boundary placement as a discrete decision process optimized via Group Relative Policy Gradient (GRPG), ReinPatch bypasses the need for continuous relaxations and performs dynamic patching policy optimization in a natural manner. Moreover, our method allows strict enforcement of a desired compression rate, freeing the downstream backbone to scale efficiently, and naturally supports multi-level hierarchical modeling. We evaluate ReinPatch on time-series forecasting datasets, where it demonstrates compelling performance compared to state-of-the-art data-driven patching strategies. Furthermore, our detached design allows the patching module to be extracted as a standalone foundation patcher, providing the community with visual and empirical insights into the segmentation behaviors preferred by a purely performance-driven neural patching strategy.

Related papers

• Adaptive Latent Trajectory Anchoring for Action Segmentation Dataset Condensation [53.60856261088647]
  This paper proposes to shift the condensation paradigm from optimization-based inversion to deterministic latent mapping.<n>We represent action segments as continuous trajectories anchored by sparse latent points in the noise manifold.<n>Our framework significantly outperforms state-of-the-art methods in segmentation performance across common datasets.
  arXiv  Detail & Related papers  (2026-07-10T04:00:26Z)
• Planning-aligned Token Compression for Long-Context Autonomous Driving [95.59023657139208]
  We propose a planning-aligned working memory framework built on conditional VQ-VA.<n>Compression is conditioned on both historical trajectory and a learned planning intent.<n>We evaluate on high-signal dynamic scenarios where historical context is most critical for behavior correctness.
  arXiv  Detail & Related papers  (2026-06-05T17:16:21Z)
• Power Reinforcement Post-Training of Text-to-Image Models with Super-Linear Advantage Shaping [66.25536973294726]
  Post-training methods for text-to-image (T2I) models are prone to reward hacking.<n>Super-Linear Advantage Shaping (SLAS) reshapes the local policy space.<n>SLAS consistently surpasses the DanceGRPO baseline across multiple backbones and benchmarks.
  arXiv  Detail & Related papers  (2026-05-11T17:59:25Z)
• Perceive, Route and Modulate: Dynamic Pattern Recalibration for Time Series Forecasting [68.53528085362642]
  Current deep forecasting models rely on fixed weight matrices applied uniformly to all temporal tokens.<n>We introduce Dynamic Pattern Recalibration, a backbone-agnostic mechanism that resolves this via token-level recalibration.<n>DPRNet achieves competitive performance across 12 benchmarks, validating dynamic recalibration against macroscopic parameter scaling.
  arXiv  Detail & Related papers  (2026-05-07T14:12:47Z)
• Temporal Patch Shuffle (TPS): Leveraging Patch-Level Shuffling to Boost Generalization and Robustness in Time Series Forecasting [13.90015544663242]
  We propose Temporal Patch Shuffle (TPS), a simple and model-agnostic data augmentation method for forecasting.<n>TPS extracts overlapping temporal patches, shuffles a subset of patches using variance-based ordering as a conservative, and reconstructs the sequence by averaging overlapping regions.<n>This design increases sample diversity while preserving forecast-consistent local temporal structure.
  arXiv  Detail & Related papers  (2026-04-10T07:45:02Z)
• TimeSqueeze: Dynamic Patching for Efficient Time Series Forecasting [46.671658381710785]
  TimeSqueeze is a dynamic patching mechanism that adaptively selects patch boundaries within each sequence based on local signal complexity.<n>TimeSqueeze attains up to 20x faster convergence and 8x higher data efficiency compared to equivalent point-token baselines.
  arXiv  Detail & Related papers  (2026-03-11T22:38:13Z)
• StepVAR: Structure-Texture Guided Pruning for Visual Autoregressive Models [98.72926158261937]
  We propose a training-free token pruning framework for Visual AutoRegressive models.<n>We employ a lightweight high-pass filter to capture local texture details, while leveraging Principal Component Analysis (PCA) to preserve global structural information.<n>To maintain valid next-scale prediction under sparse tokens, we introduce a nearest neighbor feature propagation strategy.
  arXiv  Detail & Related papers  (2026-03-02T11:35:05Z)
• SEER: Transformer-based Robust Time Series Forecasting via Automated Patch Enhancement and Replacement [9.482558107303646]
  This study proposes a robust time series forecasting framework called SEER.<n>In real-world time series, there are often low-quality issues during data collection, such as missing values, distribution shifts, anomalies and white noise.<n>We introduce a Learnable Patch Replacement Module, which enhances forecasting robustness and model accuracy through a two-stage process.
  arXiv  Detail & Related papers  (2026-01-31T08:12:24Z)
• EntroPE: Entropy-Guided Dynamic Patch Encoder for Time Series Forecasting [50.794700596484894]
  We propose EntroPE (Entropy-Guided Dynamic Patch), a novel, temporally informed framework that dynamically detects transition points via conditional entropy.<n>This preserves temporal structure while retaining the computational benefits of patching.<n> Experiments across long-term forecasting benchmarks demonstrate that EntroPE improves both accuracy and efficiency.
  arXiv  Detail & Related papers  (2025-09-30T12:09:56Z)
• DeCoP: Enhancing Self-Supervised Time Series Representation with Dependency Controlled Pre-training [39.30046923897652]
  We propose a Dependency Controlled Pre-training framework that explicitly models dynamic, multi-scale dependencies by simulating evolving inter-patch dependencies.<n>DeCoP achieves state-of-the-art results on ten datasets with lower computing resources, improving MSE by 3% on ETTh1 over PatchTST using only 37% of the FLOPs.
  arXiv  Detail & Related papers  (2025-09-18T05:44:06Z)
• Controllable Patching for Compute-Adaptive Surrogate Modeling of Partial Differential Equations [5.248019942004983]
  We introduce two lightweight modules that enable dynamic patch size control inference in patch based models without retraining or accuracy loss.<n>To our knowledge, this is the first framework to enable inference-time patch-size tunability in patch-based PDE surrogates.
  arXiv  Detail & Related papers  (2025-07-12T12:16:04Z)
• Enhancing Robustness of Vision-Language Models through Orthogonality Learning and Self-Regularization [77.62516752323207]
  We introduce an orthogonal fine-tuning method for efficiently fine-tuning pretrained weights and enabling enhanced robustness and generalization. A self-regularization strategy is further exploited to maintain the stability in terms of zero-shot generalization of VLMs, dubbed OrthSR. For the first time, we revisit the CLIP and CoOp with our method to effectively improve the model on few-shot image classficiation scenario.
  arXiv  Detail & Related papers  (2024-07-11T10:35:53Z)
• Towards Continual Learning Desiderata via HSIC-Bottleneck Orthogonalization and Equiangular Embedding [55.107555305760954]
  We propose a conceptually simple yet effective method that attributes forgetting to layer-wise parameter overwriting and the resulting decision boundary distortion. Our method achieves competitive accuracy performance, even with absolute superiority of zero exemplar buffer and 1.02x the base model.
  arXiv  Detail & Related papers  (2024-01-17T09:01:29Z)

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.