The paper introduces PACEvolve, a framework that improves self‑evolving agents powered by Large Language Models by addressing their tendency to become trapped in local contexts and repeat flawed hypotheses. It does so through three techniques: Hierarchical Context Management to prune memory, Momentum‑Based Backtracking to escape local minima, and a self‑adaptive Collaborative Evolution policy to balance refinement and knowledge transfer. These methods enable the agents to maintain a global view of search momentum and achieve state‑of‑the‑art results on complex evolutionary benchmarks.
By Minghao Yan, Bo Peng, Benjamin Coleman, Ziqi Chen, Zhouhang Xie, Shuo Chen, Zhankui He, Noveen Sachdeva, Isabella Ye, Weili Wang, Chi Wang, Ed H. Chi, Fernando Pereira, Wang-Cheng Kang, Derek Zhiyuan Cheng, Beidou Wang
EvoTreeNAD is a genealogy‑guided evolutionary algorithm that autonomously discovers neural architectures without a predefined seed or search space. Starting from an empty root, it builds a persistent genealogy where each node represents a complete architecture; top‑percentile values from nodes and descendants steer lineage selection. The method combines an Idea Agent that proposes variants and a Code Agent that implements them, with theoretical analysis showing stationary variation regimes and empirical results demonstrating superior performance on CIFAR‑10/100 and MedMNIST‑v2 tasks.
By Lishan Yu, Derek Jiu, Qizhen Lan, Xiaoqian Jiang
arXiv:2602. 13769v3 Announce Type: replace Abstract: Automating heuristic design in complex, experiment-driven domains requires more than iterative mutation of solution algorithms.
By Qi Liu, Ruochen Hao, Can Li, Wanjing Ma
Successful mutation strategies in evolutionary code search may contain reusable knowledge that is useful beyond a single run, and in some cases may transfer across related tasks and domains. However, existing LLM-driven evolutionary frameworks largely discard such knowledge, repeatedly rediscovering similar ideas and limiting opportunities for cross-run and cross-task learning.
OR-Agent is a multi‑agent research framework that automates heuristic design for optimization problems by structuring heuristic search as a tree‑based workflow with explicit hypothesis generation and systematic backtracking. It introduces a hierarchical, optimization‑inspired reflection system that uses short‑term reflections as verbal gradients, long‑term reflections as verbal momentum, and memory compression as semantic weight decay to guide research dynamics. Experiments on classical combinatorial optimization tasks and simulation‑based cooperative driving scenarios show that OR‑Agent outperforms strong evolutionary search baselines, with all code and data publicly available.
By Qi Liu, Ruochen Hao, Can Li, Wanjing Ma
arXiv:2608. 10795v1 Announce Type: new Abstract: Successful mutation strategies in evolutionary code search may contain reusable knowledge that is useful beyond a single run, and in some cases may transfer across related tasks and domains.
By Viktor Volkov, Valentin Khrulkov, Andrey V. Galichin, Danil Sivtsov, Nikita Glazkov, Olga Volkova, Konstantin Pchelin, Iaroslav Bespalov, Dmitry V. Dylov, Petr Anokhin, Ivan Oseledets
The paper introduces ScholarEvolve, a framework that evolves the software harness of language agents by automatically incorporating insights from recent research papers. It organizes harness improvements into functional modules, uses topic modeling to identify distinct strategies, and evaluates combinations to boost task performance. Experiments show significant gains on AppWorld and Tau2-Bench, raising Qwen3.5-27B completion rates from 49.6% to 63.6% and GPT-5.4-mini pass@1 from 72.7% to 81.9%.
By Jingbo Yang, Kwei-Herng Lai, Xiaowen Wang, Yaar Harari, Evgeniy Gabrilovich, Shiyu Chang
arXiv:2606. 29082v1 Announce Type: cross Abstract: Would experience designing faster GPU kernels also help close in on a long-standing open mathematical conjecture?
By Young-Jun Lee, Seungone Kim, Minki Kang, Alistair Cheong Liang Chuen, Zerui Chen, Seungho Han, Taehee Jung, Dongyeop Kang
The paper introduces Agentic Reasoning for Tree Search (ARTS), a method that uses a reasoning language model to navigate the hypothesis‑experiment space in scientific discovery. Unlike traditional approaches that conflate hypothesis quality with execution quality and prune search logs, ARTS evaluates prior execution logs to distinguish implementation failures from poor hypotheses and selects the next hypothesis to pursue. By employing test‑time training to embed search‑tree knowledge into model weights, ARTS achieves a 15.3% relative improvement over leading algorithms on 22 benchmark tasks and enables smaller models like Qwen3‑4B to match or exceed the performance of larger closed‑source models at lower inference cost.
By Gurusha Juneja, Arnav Kumar Jain, Deepak Nathani, William Yang Wang, Xin Eric Wang
arXiv:2609.38349v1 Announce Type: cross
Abstract: Modern agentic systems combine an AI model with a harness that controls execution and environmental interactions. Harness design strongly affects lon...
By Prithwish Jana, Mononito Goswami, Hao Liu, Xinyu Li, Langlin Huang, Zhehui Huang, Zhishen Huang, Patrick Bl\"obaum, Anoop Deoras, Purak Jain, Nikos Kanakaris, Sahika Genc
The paper introduces Strategy Accumulation and Guided Execution (SAGE), a two-stage framework that makes automated fine-tuning of large language models cumulative. In the first stage, a multi-agent pipeline uses Monte Carlo Tree Search to explore training strategies while a Distillation Agent records task-specific insights and cross-task confidence scores into a structured repository. In the second stage, SAGE retrieves relevant experience from this repository to guide training on new tasks, achieving a 12.4‑percentage‑point improvement over a baseline pipeline without accumulated experience on nine unseen tasks.
By Haoran Zhao, Wei Du, Dingwen Yang, Jixuan Huang, Junlin Shang, Lingyong Fang, Ya Guo, Tao Gui, Qi Zhang, Xuanjing Huang
The paper demonstrates that large language models can generate executable procedural content generators, enabling direct search over generator programs rather than individual levels. Using Sokoban, Zelda, Dangerous Dave, and Lode Runner, the authors evolve complete Python generators via language‑model mutation and crossover, and introduce Continual Abstraction Discovery (CAD) to extract reusable primitives into a run‑specific helper module. Experiments show that CAD consistently improves mean final best fitness across all domain and API comparisons, with learned libraries being adopted by subsequent programs and repeatedly rediscovering useful utilities.
By Matthew Siper, Ahmed Khalifa, Julian Togelius