AI Paper Atlas

Research map / Multi-agent systems and games / LLM agents

AlphaAgentEvo: Evolution-Oriented Alpha Mining via Self-Evolving Agentic Reinforcement Learning

Ziyi Tang, Xuexiong Yin, Weixing Chen, Zechuan Chen, Yongsen Zheng, Wenxuan Ye, Keze Wang, Liang Lin

ICLR 2026 · Multi-agent systems and games › LLM agents · approach: Agent data synthesis and behavior studies

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes AlphaAgentEvo, a self-evolving agentic reinforcement learning approach for alpha factor mining that aims to evolve alphas using natural-language instructions and insights from failed attempts.

Problem

Alpha mining for predictive financial alpha factors from a vast, noisy search space; existing methods such as genetic programming cannot interpret natural-language instructions and fail to learn from unsuccessful attempts, hindering systematic alpha evolution.

Approach

Frames alpha mining as a self-evolving agentic reinforcement learning process so that the agent learns from unsuccessful attempts and interprets natural-language instructions to evolve alphas, in contrast to genetic programming.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
train Qwen3-4B-thinking with self-evolving agentic RLQwen3-4B-thinking · AlphaEvo500 HS300pass@3: 0.36 -> 0.97Improved
train Qwen3-1.7B with agentic RLQwen3-1.7B · AlphaEvo500 HS300pass@3: 0.08 -> 0.77Improved
use hierarchical reward ARL instead of ToolRLQwen3-4B · AlphaEvo500 HS300pass@5: 0.81 -> 0.97Improved
use AlphaAgentEvo-4B instead of GEPA with GPT-5-minievolution framework · AlphaEvo500 CSI500pass@3: 0.86 -> 0.93Improved
add exploration reward (ablation)AlphaAgentEvo-4B · AlphaEvo500pass@3: 0.54 -> 0.65Improved
add consistency (direction-aware) reward (ablation)AlphaAgentEvo-4B · AlphaEvo500pass@3: 0.51 -> 0.65Improved
add exploration reward (ablation)AlphaAgentEvo-4B · Alpha158pass@3: 0.513 -> 0.581Improved
add consistency reward (ablation)AlphaAgentEvo-4B · Alpha158pass@3: 0.510 -> 0.581Improved
use ARL-trained agent instead of genetic programmingalpha evolution · Alpha158 (2024-2025)pass@5: 0.132 -> 0.994Improved
increase GP offspring per generationGP · Alpha158pass@3: 4 offspring 0.003 -> 20 offspring 0.125 -> 50 offspring 0.094 (2024-2025)Mixed
ARL training compared with untrained base model on diversityAlphaAgentEvo-4B · top-20 alpha similaritymaximum similarity: 0.600 -> 0.263 (vs Qwen3-4B)Improved

Experimental setup

Baselines: genetic programming

Builds on: genetic programming

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy