Automated Design of Agentic Systems
🤖 AI Summary
Current general-purpose agent systems rely heavily on manual design, hindering scalability and adaptability.
Method: This paper introduces an automated agent architecture design paradigm. We formally define “Automated Design of Agent Systems” (ADAS) as a new research direction and propose a Turing-complete, code-based agent representation that enables automatic invention of prompts, tool invocations, workflows, and composition mechanisms. Our approach integrates meta-agent search, archive-guided iterative program synthesis, and a multi-domain evaluation framework spanning programming, scientific reasoning, and mathematics—enabling end-to-end agent architecture discovery.
Contribution/Results: (1) We establish the first systematic research agenda for ADAS; (2) we introduce the first programmable and evolvable agent representation; and (3) our method consistently discovers novel agent architectures that surpass state-of-the-art performance across diverse domains, demonstrating strong robustness and cross-task, cross-model transferability.
Application Category
📝 Abstract
Researchers are investing substantial effort in developing powerful general-purpose agents, wherein Foundation Models are used as modules within agentic systems (e.g. Chain-of-Thought, Self-Reflection, Toolformer). However, the history of machine learning teaches us that hand-designed solutions are eventually replaced by learned solutions. We describe a newly forming research area, Automated Design of Agentic Systems (ADAS), which aims to automatically create powerful agentic system designs, including inventing novel building blocks and/or combining them in new ways. We further demonstrate that there is an unexplored yet promising approach within ADAS where agents can be defined in code and new agents can be automatically discovered by a meta agent programming ever better ones in code. Given that programming languages are Turing Complete, this approach theoretically enables the learning of any possible agentic system: including novel prompts, tool use, workflows, and combinations thereof. We present a simple yet effective algorithm named Meta Agent Search to demonstrate this idea, where a meta agent iteratively programs interesting new agents based on an ever-growing archive of previous discoveries. Through extensive experiments across multiple domains including coding, science, and math, we show that our algorithm can progressively invent agents with novel designs that greatly outperform state-of-the-art hand-designed agents. Importantly, we consistently observe the surprising result that agents invented by Meta Agent Search maintain superior performance even when transferred across domains and models, demonstrating their robustness and generality. Provided we develop it safely, our work illustrates the potential of an exciting new research direction toward automatically designing ever-more powerful agentic systems to benefit humanity.
Problem
Research questions and friction points this paper is trying to address.
Automating the design of agentic systems using learned solutions.
Exploring meta agents to program and discover new agentic systems.
Demonstrating superior performance and transferability of automatically designed agents.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Automated Design of Agentic Systems (ADAS)
Meta Agent Search algorithm
Turing Complete programming for agent discovery
🔎 Similar Papers
No similar papers found.
