The Resolution Hypothesis: Agent Evolution's Key Is Identifying Improvements
The Resolution Hypothesis: Agent Evolution's Key Is Identifying Improvements
The Resolution Hypothesis suggests that the evolution of agents stagnates due to noise evaluation being unable to reliably identify improvements, rather than the failure of methods to propose better candidates. In experiments with seven evolutionary methods and six models (over 10M large language model calls), it was found that candidate-level selection mechanisms cannot distinguish between beneficial and harmful candidates, and the ranking differences at the method level are smaller than the differences when the same method runs independently, indicating that the bottleneck in evolution lies in improvement identification. This requires us to shift towards development and evaluation methods that make real improvements detectable and accumulable.
Abstract
Agent evolution is often bottlenecked by noisy evaluations that cannot reliably identify better candidate artifacts or rank methods, so progress requires making true improvements detectable and accumulable.
Agent evolution follows a simple loop: a proposer writes candidate artifacts (prompts, skills, memories) and selection accepts those that score higher. This loop has shown promise, yet practitioners face a puzzle: methods trade wins across papers, and evolution often fails to bootstrap on certain tasks. We trace this puzzle to the Resolution Hypothesis: agent evolution stalls not because methods fail to propose better artifacts, but because noisy evaluations cannot reliably identify them. Across seven evolution methods and six models (over 10M LLM calls), we find support for the hypothesis at both levels where evaluation scores are used. At the candidate level (choosing artifacts), helpful candidate artifacts exist, but selection is no better than random at telling them from harmful ones, so errors accumulate across rounds. At the method level (ranking methods), typical margins of a few percentage points are smaller than the difference between two independent runs of the same method, so published rankings often rest on luck. Together, these findings indicate that agent evolution is often bottlenecked by reliably identifying improvements. This calls for a shift in how we develop and evaluate self-improving agents, from designing ever more candidate improvement systems to making true improvements detectable and accumulable.
View arXiv page View PDF Project page Add to collection
Models citing this paper 0
No model linking this paper
Cite arxiv.org/abs/2611.23068 in a model README.md to link it from this page.
Datasets citing this paper 0
No dataset linking this paper
Cite arxiv.org/abs/2611.23068 in a dataset README.md to link it from this page.
Spaces citing this paper 0
No Space linking this paper
Cite arxiv.org/abs/2611.23068 in a Space README.md to link it from this page.
Collections including this paper 0
No Collection including this paper
Add this paper to a collection to link it from this page.
Source: Hugging Face Daily Papers · huggingface.co