Compound AI Systems Optimization: A Survey of Methods, Challenges, and Future Directions

TL;DR

This survey reviews recent methods, challenges, and future directions in optimizing complex compound AI systems, emphasizing the integration of multiple components and the emergence of natural language feedback techniques.

Contribution

It formalizes compound AI system optimization, classifies existing methods, and highlights open challenges and future research directions.

Findings

Traditional methods like supervised fine-tuning and reinforcement learning are foundational.

Natural language feedback offers promising new optimization approaches.

The survey provides a comprehensive classification and analysis of recent progress.

Abstract

Recent advancements in large language models (LLMs) and AI systems have led to a paradigm shift in the design and optimization of complex AI workflows. By integrating multiple components, compound AI systems have become increasingly adept at performing sophisticated tasks. However, as these systems grow in complexity, new challenges arise in optimizing not only individual components but also their interactions. While traditional optimization methods such as supervised fine-tuning (SFT) and reinforcement learning (RL) remain foundational, the rise of natural language feedback introduces promising new approaches, especially for optimizing non-differentiable systems. This paper provides a systematic review of recent progress in optimizing compound AI systems, encompassing both numerical and language-based techniques. We formalize the notion of compound AI system optimization, classify existing methods along several key dimensions, and highlight open research challenges and future directions in this rapidly evolving field. A list of surveyed papers is publicly available at https://github.com/MiuLab/AISysOpt-Survey.