Large Language Models (LLMs) have not only revolutionized natural language processing but also extended their prowess to various domains, marking a significant stride towards artificial general intelligence. The interplay between LLMs and Evolutionary Algorithms (EAs), despite differing in objectives and methodologies, share a common pursuit of applicability in complex problems.

• LLMs provide rich domain knowledge for informed EA searches.
• EAs offer an optimization framework under black-box settings to enhance LLMs.
• A synergistic approach leads to the application of EAs in neural architecture search, code generation, and more.
• A roadmap and challenges for integrating LLMs and EAs in future research.
• The first comprehensive review discussing the potential of LLMs in EA research.

This paper is a cornerstone for those interested in the intersection of LLMs and evolutionary computation. It presents a wealth of opportunities where LLMs can augment EAs to solve more complex tasks, and vice versa. The potential for LLMs to contribute to a diverse array of generation tasks, including code generation, is particularly intriguing. This symbiotic relationship could spearhead advancements in AI and give rise to more intelligent and adaptable systems. Read the full paper here.