Automated Machine Learning: New Advances on Bayesian Optimization
31 Mar 2020 Tuesday, 03:00 PM to 04:30 PM
Executive Classroom, COM2-04-02
COM2 Level 4
Executive Classroom, COM2-04-02
Abstract:
Recent advances in Bayesian optimization (BO) have delivered a promising suite of tools for optimizing an unknown expensive to evaluate black-box objective function with a finite budget of evaluations. A significant advantage of BO is its general formulation: BO can be utilized to optimize any black-box objective function, since it does not make assumptions on the function's properties. As a result, BO algorithms have been applied in a wide range of applications like automated machine learning, robotics or environmental monitoring, among others. Furthermore, the general formulation makes BO attractive for deployment in new applications. However, potential new applications can have additional requirements not satisfied by the classical BO setting. In this thesis we aim to address some of these requirements in order to scale up BO technology for the practical use in new real-world applications.
Firstly, this thesis proposal presents the outsourced-Gaussian process-upper confidence bound (O-GP-UCB) algorithm, which is the first algorithm for privacy-preserving BO in the outsourced setting with a provable performance guarantee. We consider the outsourced setting where the entity holding the dataset and the entity performing BO are represented by different parties, and the dataset cannot be released non-privately. For example, a hospital holds a dataset of sensitive medical records and outsources the BO task on this dataset to an industrial AI company. The key idea of our approach is to make the BO performance of our algorithm similar to that of non-private GP-UCB run using the original dataset, which is achieved by using a random projection-based transformation that preserves both privacy and the pairwise distances between inputs. Our main theoretical contribution is to show that a regret bound similar to that of the standard GP-UCB algorithm can be established for our O-GP-UCB algorithm. We empirically evaluate the performance of our O-GP-UCB algorithm with synthetic and real-world datasets.
Secondly, we present a principled multi-staged Bayesian sequential decision algorithm for nonmyopic adaptive BO that, in particular, exploits macro-actions inherent to the structure of several real-world task environments/applications for scaling up to a further lookahead to match up to a larger available budget. To achieve this, we first generalize GP-UCB to a new acquisition function defined with respect to a nonmyopic adaptive macro-action policy, which, unfortunately, is intractable to be optimized exactly due to an uncountable set of candidate outputs. The key novel contribution of our work here is to show that it is in fact possible to solve for a nonmyopic adaptive epsilon-Bayes-optimal macro-action BO (epsilon-Macro-BO) policy given an arbitrarily user-specified loss bound epsilon via stochastic sampling in each planning stage which requires only a polynomial number of samples in the length of macro-actions. To perform nonmyopic adaptive BO in real time, we then propose an asymptotically optimal anytime variant of our epsilon-Macro-BO algorithm with a performance guarantee. Empirical evaluation on synthetic and real-world datasets shows that our proposed approach outperforms existing state-of-the-art algorithms.
As continuation of this thesis proposal, another work applying BO to adversarial learning will be done in the next few months and presented in the final thesis.