Projective Preferential Bayesian Optimization

Bayesian optimization is an effective method for finding extrema of a black-box function. We propose a new type of Bayesian optimization for learning user preferences in high-dimensional spaces. The central assumption is that the underlying objective function cannot be evaluated directly, but instead a minimizer along a projection can be queried, which we call a projective preferential query. The form of the query allows for feedback that is natural for a human to give, and which enables interaction. This is demonstrated in a user experiment in which the user feedback comes in the form of optimal position and orientation of a molecule adsorbing to a surface. We demonstrate that our framework is able to find a global minimum of a high-dimensional black-box function, which is an infeasible task for existing preferential Bayesian optimization frameworks that are based on pairwise comparisons.

[1]  Donald R. Jones,et al.  Efficient Global Optimization of Expensive Black-Box Functions , 1998, J. Glob. Optim..

[2]  F. Mosteller.  Remarks on the method of paired comparisons: I. The least squares solution assuming equal standard deviations and equal correlations , 1951 .

[3]  Nando de Freitas,et al.  Active Preference Learning with Discrete Choice Data , 2007, NIPS.

[4]  H. Simon,et al.  Reason in Human Affairs. , 1984 .

[5]  Jukka Corander,et al.  Bayesian inference of atomistic structure in functional materials , 2017, npj Computational Materials.

[6]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[7]  Matthias Scheffler,et al.  Ab initio molecular simulations with numeric atom-centered orbitals , 2009, Comput. Phys. Commun..

[8]  G. Debreu.  Mathematical Economics: Representation of a preference ordering by a numerical function , 1983 .

[9]  Nando de Freitas,et al.  A Bayesian interactive optimization approach to procedural animation design , 2010, SCA '10.

[10]  Takeo Igarashi,et al.  Sequential line search for efficient visual design optimization by crowds , 2017, ACM Trans. Graph..

[11]  D. McFadden.  Econometric Models of Probabilistic Choice , 1981 .

[12]  Wei Chu,et al.  Preference learning with Gaussian processes , 2005, ICML.

[13]  Hal S. Stern,et al.  A continuum of paired comparisons models , 1990 .

[14]  F. Mosteller.  Remarks on the method of paired comparisons: I. The least squares solution assuming equal standard deviations and equal correlations , 1951 .

[15]  J. Kadane.  Structural Analysis of Discrete Data with Econometric Applications , 1984 .

[16]  James C. Spall,et al.  Cyclic Seesaw Process for Optimization and Identification , 2012, J. Optim. Theory Appl..

[17]  Neil D. Lawrence,et al.  Preferential Bayesian Optimization , 2017, ICML.

[18]  L. Thurstone.  A law of comparative judgment. , 1994 .

[19]  Matthew W. Hoffman,et al.  Predictive Entropy Search for Efficient Global Optimization of Black-box Functions , 2014, NIPS.

[20]  Eric Brochu,et al.  Interactive Bayesian optimization : learning user preferences for graphics and animation , 2010 .

[21]  Andreas Krause,et al.  Adaptive and Safe Bayesian Optimization in High Dimensions via One-Dimensional Subspaces , 2019, ICML.

[22]  David C. Parkes,et al.  Preference Elicitation For General Random Utility Models , 2013, UAI.

[23]  L. Weatherford,et al.  Implementation of MNL choice models in the passenger origin-destination simulator (PODS) , 2015 .