Parallelised Bayesian Optimisation via Thompson Sampling

By Kirthevasan Kandasamy, Akshay Krishnamurthy, Jeff Schneider, Barnab'as P'oczos. 2018

291 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

BANANAS: Bayesian Optimization with Neural Architectures for Neural Architecture Search (2019, methodology): “Note also that both TS and ITS have advantages when running parallel experiments, since they are stochastic acquisition functions that can be directly applied in the batch BO setting (Kandasamy et al. 2018a).”
Benchmark and Survey of Automated Machine Learning Frameworks (2019, background): “Pending candidate evaluations can be either ignored—if sampling a new candidate depends on a stochastic process (Bergstra et al., 2011; Kandasamy et al., 2018)— or imputed with a constant (Ginsbourger et al., 2010b) or simulated performance (Ginsbourger et al., 2010a; Snoek et al., 2012; Desautels…”
A Tutorial on Thompson Sampling (2017, background): “, 2015), hyperparameter tuning (Kandasamy et al., 2018), and arcade games”
Machine learning applications in drug development (2019, methodology): “Drug discovery problems have also motivated research in blackbox optimization, mostly using Bayesian Optimization (BO), see for instance the following referenced examples [53,54].”
Recent Advances in Bayesian Optimization (2022, background): “achieves a balance between exploitation and exploration [123].”
Scalable Global Optimization via Local Bayesian Optimization (2019, methodology): “[26] have shown that the TS acquisition function is efficient for batch acquisition with a single global surrogate model.”
Constrained Bayesian Optimization with Noisy Experiments (2017, methodology): “Recently the classic Thompson sampling algorithm (Thompson, 1933) has been applied to GP Bayesian optimization (Hernández-Lobato et al., 2017; Kandasamy et al., 2018).”
Bayesian Optimization for Categorical and Category-Specific Continuous Inputs (2019, unknown): “The key difference is that we get probabilistic regret bounds holding with high probability. analyses; and (3) TS is shown to achieve competitive performance in practice due to using a complete distribution for exploration/exploitation trade-off (Kandasamy et al. 2018).”
Bayesian reaction optimization as a tool for chemical synthesis (2021, background): “…as a Gaussian process represents a distribution over functions, one can draw and maximize candidate models that fit the data where information is available and vary according to function shape and estimated variance in areas where little is known—a process called Thompson sampling 39,40 .”