---
format: "aidr-story-markdown/v1"
id: "2c378455d08ebf2bd636919a9518cd5036df2d45d270720e62964e416ce3e6bf"
canonical_url: "https://aidr.today/2c378455?lang=en"
title: "AllocFlow: End-to-End Portfolio Trading Action Generation from Raw Market Data"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-17T00:00:00.000Z"
category: "Research"
topics: ["agent"]
source_urls: ["https://huggingface.co/papers/2611.19937"]
summary: "AllocFlow is an end-to-end flow-matching framework that generates joint portfolio allocations directly from raw multi-asset market histories. It uses a cross-asset flow Transformer and a two-stage training scheme involving strategy-supervised imitation and group-relative reinforcement learning."
---

# AllocFlow: End\-to\-End Portfolio Trading Action Generation from Raw Market Data

> [Open the canonical story](<https://aidr.today/2c378455?lang=en>)

**Published:** 2026-11-17T00:00:00.000Z
**Category:** Research
**Topics:** agent

## Summary

AllocFlow is an end\-to\-end flow\-matching framework that generates joint portfolio allocations directly from raw multi\-asset market histories\. It uses a cross\-asset flow Transformer and a two\-stage training scheme involving strategy\-supervised imitation and group\-relative reinforcement learning\.

## Sources

- [Story source](<https://huggingface.co/papers/2611.19937>)

