Paper

Coverage-Based Designs Improve Sample Mining and Hyper-Parameter Optimization

Sampling one or more effective solutions from large search spaces is a recurring idea in machine learning, and sequential optimization has become a popular solution. Typical examples include data summarization, sample mining for predictive modeling and hyper-parameter optimization. Existing solutions attempt to adaptively trade-off between global exploration and local exploitation, wherein the initial exploratory sample is critical to their success. While discrepancy-based samples have become the de facto approach for exploration, results from computer graphics suggest that coverage-based designs, e.g. Poisson disk sampling, can be a superior alternative. In order to successfully adopt coverage-based sample designs to ML applications, which were originally developed for 2-d image analysis, we propose fundamental advances by constructing a parameterized family of designs with provably improved coverage characteristics, and by developing algorithms for effective sample synthesis. Using experiments in sample mining and hyper-parameter optimization for supervised learning, we show that our approach consistently outperforms existing exploratory sampling methods in both blind exploration, and sequential search with Bayesian optimization.

Results in Papers With Code
(↓ scroll down to see all results)