Saved in:
Bibliographic Details
Main Authors: Ye, Liqin, Yin, Yanbin, Galarnyk, Michael, Heng, Yuzhao, Chava, Sudheer, Zhang, Chao
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.11666
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913116038627328
author Ye, Liqin
Yin, Yanbin
Galarnyk, Michael
Heng, Yuzhao
Chava, Sudheer
Zhang, Chao
author_facet Ye, Liqin
Yin, Yanbin
Galarnyk, Michael
Heng, Yuzhao
Chava, Sudheer
Zhang, Chao
contents The reasoning frontier of Large Language Models (LLMs) has advanced significantly through modern post-training paradigms (e.g., Reinforcement Learning from Verifiable Rewards (RLVR)). However, the efficacy of these methods remains fundamentally constrained by the diversity and complexity of the training data. One practical solution is data synthesis; yet, prevalent methods relying on unstructured mutation or exploration suffer from homogeneity collapse, failing to systematically expand the reasoning frontier. To overcome this, we propose Evoutionary Task Discovery (EvoTD), a framework that treats data synthesis as a directed search over a dual-axis manifold of Algorithmic Skills and Complexity Attributes. We introduce structured evolutionary operators to navigate this space: a Crossover operator that synthesizes novel skill compositions to enhance diversity, and a Parametric Mutation operator that scales structural constraints (e.g., input size, tree depth) to drive robust generalization. Crucially, we integrate a dynamic Zone of Proximal Development filter, ensuring tasks lie within the learnable region of the model. Empirically, EvoTD delivers substantial reasoning gains that generalize consistently across model architectures, pretraining regimes, and scales, demonstrating that structured evolutionary curricula can effectively support reasoning improvement. We release our code on https://github.com/liqinye/EvoTD.
format Preprint
id arxiv_https___arxiv_org_abs_2605_11666
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Evolutionary Task Discovery: Advancing Reasoning Frontiers via Skill Composition and Complexity Scaling
Ye, Liqin
Yin, Yanbin
Galarnyk, Michael
Heng, Yuzhao
Chava, Sudheer
Zhang, Chao
Machine Learning
Artificial Intelligence
The reasoning frontier of Large Language Models (LLMs) has advanced significantly through modern post-training paradigms (e.g., Reinforcement Learning from Verifiable Rewards (RLVR)). However, the efficacy of these methods remains fundamentally constrained by the diversity and complexity of the training data. One practical solution is data synthesis; yet, prevalent methods relying on unstructured mutation or exploration suffer from homogeneity collapse, failing to systematically expand the reasoning frontier. To overcome this, we propose Evoutionary Task Discovery (EvoTD), a framework that treats data synthesis as a directed search over a dual-axis manifold of Algorithmic Skills and Complexity Attributes. We introduce structured evolutionary operators to navigate this space: a Crossover operator that synthesizes novel skill compositions to enhance diversity, and a Parametric Mutation operator that scales structural constraints (e.g., input size, tree depth) to drive robust generalization. Crucially, we integrate a dynamic Zone of Proximal Development filter, ensuring tasks lie within the learnable region of the model. Empirically, EvoTD delivers substantial reasoning gains that generalize consistently across model architectures, pretraining regimes, and scales, demonstrating that structured evolutionary curricula can effectively support reasoning improvement. We release our code on https://github.com/liqinye/EvoTD.
title Evolutionary Task Discovery: Advancing Reasoning Frontiers via Skill Composition and Complexity Scaling
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2605.11666