The rapid advancement of large language models (LLMs) has significantly enhanced the capabilities of AI-driven agents across various tasks. However, existing agentic systems, whether based on fixed pipeline algorithms or pre-defined meta-learning frameworks, cannot search the whole agent design space due to the restriction of human-designed components, and thus might miss the globally optimal agent design. In this paper, we introduce Godel Agent, a self-evolving frame- ¨ work inspired by the Godel machine, enabling agents to recursively improve them- ¨ selves without relying on predefined routines or fixed optimization algorithms. Godel Agent leverages LLMs to dynamically modify its own logic and behavior, ¨ guided solely by high-level objectives through prompting. Experimental results on multiple domains including coding, science, and math demonstrate that implementation of Godel Agent can achieve continuous self-improvement, surpassing ¨ manually crafted agents in performance, efficiency, and generalizability
We propose Godel Agent, a self-referential framework that enables agents to recursively improve themselves, overcoming the limitations of hand-designed agents and meta-learning optimized agents. Godel Agent can dynamically modify its own logic based on high-level objectives. Experimental results demonstrate its superior performance, efficiency, and adaptability compared to traditional agents. This research lays the groundwork for a new paradigm in autonomous agent development, where LLMs, rather than human-designed constraints, define the capabilities of AI systems. Realizing this vision will require the collective efforts of the entire research community.
A Gödel machine is a hypothetical self-improving computer program that solves problems in an optimal way. It uses a recursive self-improvement protocol in which it rewrites its own code when it can prove the new code provides a better strategy.[1][2] The machine was invented by Jürgen Schmidhuber (first proposed in 2003[3]), but is named after Kurt Gödel who inspired the mathematical theories.[4]
The Gödel machine is often discussed when dealing with issues of meta-learning, also known as “learning to learn.” Applications include automating human design decisions and transfer of knowledge between multiple related tasks, and may lead to design of more robust and general learning architectures.[5] Though theoretically possible, no full implementation has been created.[6]
The Gödel machine is often compared with Marcus Hutter’s AIXI, another formal specification for an artificial general intelligence. Schmidhuber points out that the Gödel machine could start out by implementing AIXItl as its initial sub-program, and self-modify after it finds proof that another algorithm for its search code will be better.[7]
Self-Improving code generation
Several recent advances in AI systems solve problems by providing a “scaffolding” program that structures multiple calls to language models (LMs) to generate better outputs. A scaffolding program is written in a programming language such as Python. In this work, we use a language-model-infused scaffolding program to improve itself. We start with a seed “improver” that improves an input program according to a given utility function by querying an LM several times and returning the best solution. We then run this seed improver to improve itself. Across a small set of downstream tasks, the resulting improved improver generates programs with significantly better performance than its seed improver. A variety of self-improvement strategies are proposed by the language model, including beam search, genetic algorithms, and simulated annealing. Since the language models themselves are not altered, this is not full recursive self-improvement. Nonetheless, it demonstrates that a modern language model, GPT-4 in our experiments, is capable of writing code that can call itself to improve itself. We consider concerns around the development of self-improving technologies and evaluate the frequency with which the generated code bypasses a sandbox
In this work, we introduced STOP, a framework for recursively optimizing code generation using LMs as meta-optimizers. We demonstrated that LMs like GPT-4 are capable of improving code that leverages the LM itself. We found that, across a variety of algorithmic tasks, STOP generates improvers that boost the performance of downstream code. While the model does not optimize its weights or underlying architecture, this work indicates that self-optimizing LMs do not require that. However, this is itself a motivation: the capabilities of future LMs may be misunderstood if strong scaffolding strategies are not tested. Understanding how LMs can improve their scaffoldings can help researchers understand and mitigate the potential for misuse of more powerful LMs. Lastly, STOP may allow researchers to investigate techniques for mitigating undesirable self-improvement strategies.
