LLM Guided Evolution -- The Automation of Models Advancing Models

TL;DR

This paper introduces Guided Evolution (GE), a novel framework using Large Language Models to directly modify code for model evolution, improving accuracy and maintaining diversity without relying on traditional AutoML methods.

Contribution

The study presents a new LLM-guided evolutionary framework that enables autonomous model development with a self-reflective feedback loop, diverging from conventional AutoML techniques.

Findings

Achieved accuracy increase from 92.52% to 93.34% in evolved models.

Maintained model compactness while improving performance.

Demonstrated autonomous model evolution using LLM-guided mutations.

Abstract

In the realm of machine learning, traditional model development and automated approaches like AutoML typically rely on layers of abstraction, such as tree-based or Cartesian genetic programming. Our study introduces "Guided Evolution" (GE), a novel framework that diverges from these methods by utilizing Large Language Models (LLMs) to directly modify code. GE leverages LLMs for a more intelligent, supervised evolutionary process, guiding mutations and crossovers. Our unique "Evolution of Thought" (EoT) technique further enhances GE by enabling LLMs to reflect on and learn from the outcomes of previous mutations. This results in a self-sustaining feedback loop that augments decision-making in model evolution. GE maintains genetic diversity, crucial for evolutionary algorithms, by leveraging LLMs' capability to generate diverse responses from expertly crafted prompts and modulate model temperature. This not only accelerates the evolution process but also injects expert like creativity and insight into the process. Our application of GE in evolving the ExquisiteNetV2 model demonstrates its efficacy: the LLM-driven GE autonomously produced variants with improved accuracy, increasing from 92.52% to 93.34%, without compromising model compactness. This underscores the potential of LLMs to accelerate the traditional model design pipeline, enabling models to autonomously evolve and enhance their own designs.