Man-Made Heuristics Are Dead. Long Live Code Generators!
This work addresses the challenge of automating policy design for systems controllers, offering a novel approach that could reduce reliance on manual expert heuristics, though it appears incremental as it builds on existing generative models.
The paper tackles the problem of manual policy design for systems controllers by introducing PolicySmith, an automated framework that uses LLM-driven code generation to synthesize instance-optimal heuristics, resulting in discovered heuristics that outperform established baselines in web caching and generate safe policies for congestion control in the Linux kernel.
Policy design for various systems controllers has conventionally been a manual process, with domain experts carefully tailoring heuristics for the specific instance in which the policy will be deployed. In this paper, we re-imagine policy design via a novel automated search technique fueled by recent advances in generative models, specifically Large Language Model (LLM)-driven code generation. We outline the design and implementation of PolicySmith, a framework that applies LLMs to synthesize instance-optimal heuristics. We apply PolicySmith to two long-standing systems policies - web caching and congestion control, highlighting the opportunities unraveled by this LLM-driven heuristic search. For caching, PolicySmith discovers heuristics that outperform established baselines on standard open-source traces. For congestion control, we show that PolicySmith can generate safe policies that integrate directly into the Linux kernel.