• Publications
  • Influence
High Dimensional Bayesian Optimisation and Bandits via Additive Models
TLDR
Bayesian Optimisation (BO) is a technique used in optimising a $D$-dimensional function which is typically expensive to evaluate. Expand
  • 169
  • 24
  • PDF
Neural Architecture Search with Bayesian Optimisation and Optimal Transport
TLDR
We develop NASBOT, a Gaussian process based BO framework for neural architecture search and demonstrate that it outperforms other alternatives for architecture search. Expand
  • 227
  • 19
  • PDF
Multi-fidelity Bayesian Optimisation with Continuous Approximations
TLDR
We study multi-fidelity optimisation when there is access to a continuous spectrum of approximations. Expand
  • 76
  • 12
  • PDF
Gaussian Process Bandit Optimisation with Multi-fidelity Evaluations
TLDR
We develop \mfgpucb, a novel method based on upper confidence bound techniques. Expand
  • 69
  • 12
  • PDF
Additive Approximations in High Dimensional Nonparametric Regression via the SALSA
TLDR
We propose SALSA, which bridges this gap by allowing interactions between variables, but controls model capacity by limiting the order of interactions. Expand
  • 29
  • 11
  • PDF
Nonparametric Estimation of Renyi Divergence and Friends
TLDR
We consider nonparametric estimation of L2, Renyi-α and Tsallis-α divergences between continuous distributions. Expand
  • 74
  • 10
  • PDF
Parallelised Bayesian Optimisation via Thompson Sampling
TLDR
We design and analyse variations of the classical Thompson sampling (TS) procedure for Bayesian optimisation (BO) in settings where function evaluations are expensive but can be performed in parallel. Expand
  • 90
  • 9
  • PDF
Nonparametric von Mises Estimators for Entropies, Divergences and Mutual Informations
TLDR
We propose and analyse estimators for statistical functionals of one or more distributions under nonparametric assumptions and show the advantage of this approach over existing estimators both theoretically and empirically. Expand
  • 67
  • 6
  • PDF
Tuning Hyperparameters without Grad Students: Scalable and Robust Bayesian Optimisation with Dragonfly
TLDR
We present Dragonfly, an open source Python library for scalable and robust Bayesian Optimisation of expensive black box functions, which can be applied in real world settings. Expand
  • 34
  • 6
  • PDF
High Dimensional Bayesian Optimization via Restricted Projection Pursuit Models
TLDR
We generalize the existing assumption to a projected-additive assumption for Bayesian Optimization and propose a restricted-projection-pursuit GP. Expand
  • 46
  • 3
  • PDF