FlowTSFM: Turning Encoder Depth into Quantile Transport

2026-09-11

AI Summary

本文提出FlowTSFM模型，通过共享参数的单个Transformer块迭代应用和分位数流目标监督，解决传统时间序列基础模型中深度编码器中间表示无预测作用的问题。

Abstract

Encoder-based time series foundation models (TSFMs) typically rely on deep stacks of independently parameterized Transformer layers, where only the final forecast is supervised and intermediate representations have no explicit predictive role. We introduce FlowTSFM, an encoder architecture that interprets depth as a recurrent transport process: a single Transformer block is iteratively applied with shared parameters, while a quantile-flow objective supervises intermediate states along a prescribed trajectory from a prior distribution toward the final forecast. The objective combines pinball forecasting loss with path-level position matching. With only 38.8M parameters, FlowTSFM achieves competitive performance on GIFT-Eval and TIME, remaining within 1.8-4.6% MASE of stronger baselines while using approximately $3\times$ fewer parameters than a 12-layer Chronos-2 model (119.5M). Beyond accuracy, we introduce CosMean, a scale-free diagnostic measuring whether recurrent updates consistently align toward the final prediction. Under a matched intermediate-state probing protocol, FlowTSFM achieves a CosMean score of 0.919 compared with 0.350 for Chronos-2, suggesting that recurrent parameter sharing combined with path supervision is associated with substantially more structured predictive trajectories at a favorable accuracy-efficiency trade-off.

Problem

• Time Series Foundation Models
• Transformer Layers
• Intermediate Representations

Innovation

• FlowTSFM
• quantile-flow objective
• recurrent transport process
• shared parameters
• CosMean