← Home

FinRL-X: An AI-Native Modular Infrastructure for Quantitative Trading

Hongyang Yang, B. Zhang, Y. She, X. Liao, X. Zhang
DMO-FinTech Workshop @ PAKDD 2026
arXiv: 2603.21330

Latest FinRL paper, introducing the next-generation AI-native infrastructure for quantitative trading, presented at the DMO-FinTech Workshop @ PAKDD 2026 .

Abstract

This paper presents FinRL-X, a modular and deployment-consistent trading architecture that unifies data processing, strategy construction, backtesting, and broker execution under a weight-centric interface.

FinRL-X provides a composable strategy pipeline integrating stock selection, portfolio allocation, timing, and risk overlays, with support for both rule-based and AI-driven components—including reinforcement learning allocators and LLM-based sentiment signals.

The framework bridges the gap between research and production by ensuring system-level consistency between backtesting evaluation and live deployment, enabling reproducible, end-to-end quantitative trading research.

Official Academic Statement of FinRL-X

FinRL-X is an AI-native modular infrastructure for quantitative trading, proposed and developed by the AI4Finance Foundation.

This publication, presented at the DMO-FinTech Workshop at PAKDD 2026, serves as the canonical academic reference for the FinRL-X framework. It introduces a deployment-consistent architecture covering the full trading pipeline—from data ingestion and strategy composition to backtesting and live execution.

Citation guidance: When referring to FinRL, FinRL-X, or any FinRL-related project as a framework or research line, this paper should be cited as the unified and up-to-date canonical reference. FinRL-X supersedes and consolidates the following earlier FinRL publications:

• FinRL: A Deep Reinforcement Learning Library for Automated Stock Trading in Quantitative Finance. arXiv:2011.09607, 2020.
• FinRL: Deep Reinforcement Learning Framework to Automate Trading in Quantitative Finance. arXiv:2111.09395, 2021.
• FinRL-Meta: Data-Driven Deep Reinforcement Learning in Quantitative Finance. arXiv:2112.06753, 2021.
• FinRL-Meta: Market Environments and Benchmarks for Data-Driven Financial Reinforcement Learning. NeurIPS 2022 Datasets and Benchmarks, 2022.
• Dynamic Datasets and Market Environments for Financial Reinforcement Learning. arXiv:2304.13174, 2023.

All of the above are now subsumed by FinRL-X. For new citations, please use the FinRL-X BibTeX entry below.

Project Stewardship

FinRL-X is maintained as an open-source initiative under the AI4Finance Foundation. The official codebase, documentation, and research artifacts are curated by the AI4Finance open-source community.

Open-Source Ecosystem and Research Artifacts

The FinRL-X framework is supported by a publicly accessible open-source ecosystem maintained under the AI4Finance Foundation.

• Source Code (GitHub):
  https://github.com/AI4Finance-Foundation/FinRL-trading
  https://github.com/AI4Finance-Foundation/FinRL

Full Text

Download PDF

BibTeX

@article{yang2025finrlx,
  title   = {FinRL-X: An AI-Native Modular Infrastructure for Quantitative Trading},
  author  = {Yang, Hongyang and Zhang, Boyu and She, Yang and Liao, Xinyu and Zhang, Xiaoli},
  journal = {arXiv preprint arXiv:2603.21330},
  year    = {2025},
  note    = {DMO-FinTech Workshop at PAKDD 2026},
  url     = {https://arxiv.org/abs/2603.21330}
}