Abstrakti
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.
Alkuperäiskieli | Englanti |
---|---|
Otsikko | Proceedings of the 37th International Conference on Machine Learning |
Sivumäärä | 9 |
Vuosikerta | 119 |
Tila | Julkaistu - 21 marraskuuta 2020 |
OKM-julkaisutyyppi | A4 Artikkeli konferenssijulkaisuussa |
Tapahtuma | International Conference on Machine Learning - Vienna, Itävalta Kesto: 12 heinäkuuta 2020 → 18 heinäkuuta 2020 Konferenssinumero: 37 |
Julkaisusarja
Nimi | Proceedings of Machine Learning Research |
---|---|
ISSN (elektroninen) | 2640-3498 |
Conference
Conference | International Conference on Machine Learning |
---|---|
Lyhennettä | ICML |
Maa | Itävalta |
Kaupunki | Vienna |
Ajanjakso | 12/07/2020 → 18/07/2020 |