Summary of Sober: Highly Parallel Bayesian Optimization and Bayesian Quadrature Over Discrete and Mixed Spaces, by Masaki Adachi et al.
SOBER: Highly Parallel Bayesian Optimization and Bayesian Quadrature over Discrete and Mixed Spaces
by Masaki Adachi, Satoshi Hayakawa, Saad Hamid, Martin Jørgensen, Harald Oberhauser, Micheal A. Osborne
First submitted to arxiv on: 27 Jan 2023
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Numerical Analysis (math.NA); Computation (stat.CO); Machine Learning (stat.ML)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed algorithm, SOBER, addresses the challenge of scaling batch global optimization and quadrature to large sizes while maintaining diversity. By reformulating batch selection as a quadrature problem, SOBER relaxes non-convex acquisition function maximization to convex kernel recombination. This approach efficiently solves both tasks by balancing exploitative Bayesian optimization and explorative Bayesian quadrature. Experimental results demonstrate that SOBER outperforms 11 competitive baselines on 12 synthetic and real-world tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary SOBER is a new way to do optimization and quadrature when we have lots of things to try at once. It’s like a recipe for finding the best answer or making an estimate. Right now, other methods don’t work well when there are many options, but SOBER can handle that. The key idea is to change how we choose which option to try next. Instead of looking at just one thing, we look at lots of things together. This makes it easier and more efficient to find the best answer. |
Keywords
* Artificial intelligence * Optimization
