Problem
Research questions and friction points this paper is trying to address.
rare event sampling
distribution matching
evolutionary algorithms
generative models
sample efficiency
Distribution Matching Evolutionary Algorithms for Rare Event Sampling
🤖 AI Summary
This study addresses the high computational cost of gradient-based optimization and the inaccessibility of closed-source models in rare event sampling. For the first time, it establishes a theoretical connection between evolutionary algorithms (EAs) and Markov Chain Monte Carlo (MCMC) methods by reinterpreting EAs as approximate MCMC processes. Building upon this insight, the Distribution Matching Evolutionary (DME) algorithm is proposed to enable efficient sampling from a global target distribution without updating model weights. Experimental results demonstrate that DME achieves significantly superior sample efficiency compared to existing baselines in sampling-intensive tasks. This work introduces a novel paradigm for rare event sampling under black-box model settings.
📝 Abstract
A novel discovery is one which is both useful and surprising: a generative model's output is a useful discovery if it has a low probability of being generated (it's surprising) and a high reward (it's useful). Global optimization can directly increase the probability of sampling high rewards but typically requires updating model weights. Such gradient based optimization is expensive and bars using capable closed-source models. Instead, modern search methods for discovery sacrifice the global target, and use evolutionary algorithms with local reward maximizing objectives, permitting the search to focus only on high probability samples. In this paper, we interpret various evolutionary algorithms as approximate Markov Chain Monte Carlo, an optimization-free method to sample from complex distributions. This interpretation allows developing Distribution Matching Evolutionary Algorithms (DME), a class of search methods which sample from a global target distribution without updating weights. Empirically, DME has a higher sample efficiency than existing methods on problems requiring many samples to find a solution.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Distribution Matching Evolutionary Algorithms
Rare Event Sampling
Approximate Markov Chain Monte Carlo
Weight-free Optimization
Sample Efficiency
🔎 Similar Papers
No similar papers found.
