[2511.16652] Evolution Strategies at the Hyperscale

Abstract:Evolution Strategies (ES) is a class of powerful black-box optimisation methods that are highly parallelisable and can handle non-differentiable and noisy objectives. However, naïve ES becomes prohibitively expensive at scale on GPUs due to the low arithmetic intensity of batched matrix multiplications with unstructured random perturbations. We introduce Evolution Guided GeneRal Optimisation via Low-rank Learning (EGGROLL), which improves arithmetic intensity by structuring individual perturbations as rank-$r$ matrices, resulting in a hundredfold increase in training speed for billion-parameter models at large population sizes, achieving up to 91% of the throughput of pure batch inference. We provide a rigorous theoretical analysis of Gaussian ES for high-dimensional parameter objectives, investigating conditions needed for ES updates to converge in high dimensions. Our results reveal a linearising effect, and proving consistency between EGGROLL and ES as parameter dimension increases. Our experiments show that EGGROLL: (1) enables the stable pretraining of nonlinear recurrent language models that operate purely in integer datatypes, (2) is competitive with GRPO for post-training LLMs on reasoning tasks, and (3) does not compromise performance compared to ES in tabula rasa RL settings, despite being faster.

# link_20qns2undj8.pdf ## Metadata - PDFFormatVersion=1.7 - IsLinearized=false - IsAcroFormPresent=false - IsXFAPresent=false - IsCollectionPresent=false - IsSignaturesPresent=false - Author=Bidipta Sarkar; Mattie Fellows; Juan Agustin Duque; Alistair Letcher; Antonio León Villares; Anya Sims; Clarisse Wibault; Dmitry Samsonov; Dylan Cope; Jarek Liesen; Kang Li; Lukas Seier; Theo Wolf; Uljad Berdica; Valentin Mohl; Alexander David Goldie; Aaron Courville; Karin Sevegnani; Shimon Whiteson; Jakob Nicolaus Foerster - Creator=arXiv GenPDF (tex2pdf:57610bf) - Custom.DOI=https://doi.org/10.48550/arX

saved by

• Vihan Tiwari

related reading

• A Visual Guide to Evolution Strategies | 大トロblog.otoro.net
• [2510.13786] The Art of Scaling Reinforcement Learning Compute for LLMsarxiv.org
• [2605.01172] A Theory of Generalization in Deep Learningarxiv.org
• [2507.19457] GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learningarxiv.org
• [2201.02177] Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasetsarxiv.org
• [2106.09685] LoRA: Low-Rank Adaptation of Large Language Modelsarxiv.org
• [2602.17616] Stable Asynchrony: Variance-Controlled Off-Policy RL for LLMsarxiv.org
• IsoCompute Playbook: Optimally Scaling Sampling Compute for RL Training of LLMscompute-optimal-rl-llm-scaling.github.io
• [2510.11696] QeRL: Beyond Efficiency -- Quantization-enhanced Reinforcement Learning for LLMsarxiv.org
• Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasetsmathai-iclr.github.io
• [2509.14786] Pre-training under infinite computearxiv.org
• [2509.24372] Evolution Strategies at Scale: LLM Fine-Tuning Beyond Reinforcement Learningarxiv.org