TimeTok: Granularity-Controllable Time-Series Generation via Hierarchical Tokenization

2026-05-02

AI Summary

Existing time series generative models struggle to flexibly control the temporal granularity of their outputs, making it difficult to generate sequences at user-specified levels of detail. This work proposes TimeTok, a unified framework that hierarchically tokenizes time series into ordered tokens ranging from coarse to fine granularity. TimeTok introduces an autoregressive cross-granularity generation and decoding mechanism, enabling end-to-end trainable, granularity-controllable synthesis. Notably, it is the first framework capable of generating time series at any target granularity—either from an arbitrary coarse-grained input or from scratch—by explicitly adjusting the number of token blocks to modulate detail levels. Experiments demonstrate that TimeTok achieves state-of-the-art performance on standard generation benchmarks while excelling in granularity-controllable tasks and exhibiting strong generalization and transfer capabilities across heterogeneous multi-granularity datasets.

Abstract

Time-series generative models often lack control over temporal granularity, forcing users to accept whatever granularity the model produces. To enable truly user-driven generation, we introduce TimeTok, a unified framework for Granularity-Controllable Time-Series Generation (GC-TSG), which generates time series at any target granularity from any coarser input (e.g., rough sketches) or from scratch. At the core of TimeTok is a hierarchical tokenization strategy that maps time series into an ordered sequence of tokens, from coarse to fine temporal granularity. Our autoregressive generation process operates across these granularity levels, producing token blocks that are decoded back into continuous time series. This design naturally enables GC-TSG - including standard generation - within a single framework, where controlling the number of token blocks provides explicit control over output detail. Experiments show that TimeTok excels at GC-TSG tasks while achieving state-of-the-art performance in standard generation. Furthermore, we showcase TimeTok's potential as a foundational tokenizer by training on multiple datasets with heterogeneous temporal granularities, verifying strong transferability that consistently outperforms models trained on individual datasets. To our knowledge, this is the first unified framework that covers the full generative spectrum for time series, offering a valuable foundation for models that benefit from diverse temporal granularities.

Problem

• time-series generation
• temporal granularity
• granularity control
• generative models

Innovation

• Granularity-Controllable Generation
• Hierarchical Tokenization
• Time-Series Generation
• Autoregressive Modeling
• Foundation Tokenizer