Evolutionary Ensemble of Agents
For researchers in algorithmic discovery and LLM-based optimization, EvE provides a decentralized framework that improves upon static agent systems by evolving guidance and skills, demonstrating a clear performance gain over fixed-agent baselines.
EvE organizes existing coding agents into a co-evolving system that autonomously discovered a robust rescale-then-interpolate mechanism for In-Context Operator Networks, enabling reliable example-count generalization. Controlled ablations show that stage-dependent agent adaptation is essential, and the ensemble approach uniquely avoids phase mismatch, breaking through static performance ceilings.
We introduce Evolutionary Ensemble (EvE), a decentralized framework that organizes existing, highly capable coding agents into a live, co-evolving system for algorithmic discovery. Rather than reinventing the wheel within the "LLMs as optimizers" paradigm, EvE fixes the base agent substrate and focuses entirely on evolving the cumulative guidance and skills that dictate agent behaviors. By maintaining two co-evolving populations, namely functional code solvers and agent guidance states, the system evaluates agents through a synchronous race, updating their empirical Elo ratings based on the marginal gains they contribute to the current solver state. When applied to a research bottleneck in In-Context Operator Networks (ICON), EvE autonomously discovered a robust rescale-then-interpolate mechanism that enables reliable example-count generalization. Crucially, controlled ablations reveal the absolute necessity of stage-dependent agent adaptation to navigate the shifting search landscapes of complex codebases. Compared to variants driven by a fixed initial agent or even a frozen "best-evolved" agent, EvE uniquely avoids phase mismatch, demonstrating that organizing agents into a self-revising ensemble is the fundamental driver for breaking through static performance ceilings.