News

AllocFlow Introduces End-to-End AI for Portfolio Trading

AllocFlow, a new AI framework, uses flow Transformers and reinforcement learning to optimize multi-asset portfolio trading with significant performance gains.

Thematic editorial illustration
Editorial illustration; not a product screenshot or event photograph

A new AI framework, AllocFlow, has been introduced to revolutionize portfolio management by generating joint portfolio allocations directly from raw multi-asset market data. The research, published on November 17, 2026, on arXiv (2611.19937) and supported by an open-source implementation on GitHub, demonstrates significant advancements in trading system design and performance.

Innovative End-to-End Framework

AllocFlow addresses a key limitation in traditional trading pipelines, which often separate signal forecasting from portfolio construction, potentially misaligning objectives. Instead, AllocFlow employs a cross-asset flow Transformer to process raw hourly market data and generate continuous, signed position weights for all portfolio instruments simultaneously. This approach ensures coordinated and optimized allocation decisions under shared capital constraints.

Two-Stage Training Process

The framework utilizes a two-stage training scheme to overcome inefficiencies in learning continuous allocation policies. The first stage, strategy-supervised imitation, establishes a structured distribution of candidate portfolios by mimicking a causal alpha strategy. The second stage, group-relative reinforcement learning (Flow-GRPO), refines these portfolios by sampling and scoring eight candidate books per market state over the next 24 hourly bars. This refinement process optimizes portfolio-level financial rewards while maintaining stability through a frozen reference policy.

Performance Metrics

On hourly cryptocurrency perpetual futures, AllocFlow demonstrated remarkable performance improvements. The reinforcement learning refinement increased the annualized net return of the imitation policy from 27.79% to 51.38%, while the Calmar ratio rose from 0.955 to 4.043. On six months of unseen 2026 data, AllocFlow achieved a Calmar ratio of 4.90 without retraining, surpassing the teacher strategy on all metrics.

Open-Source Availability

The AllocFlow framework is available as an open-source project on GitHub (https://github.com/allocflow/AllocFlow), allowing researchers and practitioners to explore and implement its capabilities. The project includes detailed documentation, a project page, and an explainer video outlining its methodology and results.

By integrating advanced machine learning techniques with end-to-end portfolio optimization, AllocFlow represents a significant step forward in the field of algorithmic trading and portfolio management.

Sign in to save this article Free account: build your research collection and return to read it anytime.