KiT: A Foundation Model for Financial Time-Series Forecasting using DiffusionTransformers

2026-09-28

AI Summary

This study addresses the challenges of low signal-to-noise ratios, strong market heterogeneity, and error accumulation in autoregressive models for financial candlestick prediction by proposing KiT, a novel foundation model. To our knowledge, KiT introduces the first diffusion Transformer architecture tailored for candlestick data, departing from conventional autoregressive paradigms by reformulating forecasting as a flow matching-based OHLCV trajectory generation task. Furthermore, large-scale pretraining on diverse multi-market data enhances its generalization capabilities. Extensive experiments demonstrate that, across three markets and seven temporal scales, KiT achieves an average return RankIC of 0.057 and a volatility RankIC of 0.66, comprehensively outperforming existing finance-specific and general-purpose time series foundation models.

Abstract

Financial candlestick forecasting is fundamental to quantitative investment, yet it remains exceptionally challenging due to extremely low signal-to-noise ratios and vast heterogeneity across markets and instruments. Existing approaches have largely attempted to introduce deep learning to capture hidden temporal features, but most adopt an auto-regressive formulation, which leads to error accumulation during inference. Meanwhile, general-purpose time-series foundation models are not tailored to the unique structure of k-line data and yield unsatisfactory performance on downstream candlestick forecasting tasks. To tackle these problems, we introduce KiT, a K-line Diffusion Transformer foundation model, and reformulate future prediction as conditional path generation via flow matching: given a historical context window, the model generates an ensemble of plausible future OHLCV trajectories. We pre-train KiT at multiple parameter scales on billions of candlestick bars spanning multiple markets and timescales. Across three markets and seven resolutions, KiT attains a mean return RankIC of 0.057 and a mean volatility RankIC of 0.66, leading at every timescale and outperforming both task-specific financial forecasters and general time-series foundation models. Code will be available at: https://github.com/Luciferbobo/KiT.

Problem

• Financial Time-Series Forecasting
• Candlestick Prediction
• OHLCV
• Auto-regressive Error Accumulation
• Foundation Model

Innovation

• Diffusion Transformers
• Flow Matching
• Financial Time-Series Foundation Model
• Candlestick Forecasting
• OHLCV Trajectory Generation