Black box optimization (BBO) focuses on optimizing unknown functions in high-dimensional spaces. In many applications, sampling the unknown function is expensive, imposing a tight sample budget.Ongoing work is making progress on reducing the sample budget by learning the shape/structure of the function, known as kernel learning. We propose a new method to learn the kernel of a Gaussian Process. Our idea is to create a continuous kernel space in the latent space of a variational autoencoder, and run an auxiliary optimization to identify the best kernel. Results show that the proposed method, Kernel Optimized Blackbox Optimization (KOBO), outperforms state of the art by estimating the optimal at considerably lower sample budgets. Results hold not only across synthetic benchmark functions but also in real applications. We show that a hearing aid may be personalized with fewer audio queries to the user, or a generative model could converge to desirable images from limited user ratings.

We consider the problem of personalizing audio to maximize user experience. Briefly, we aim to find a filter h*, which applied to any music or speech, will maximize the user’s satisfaction. This is a black-box optimization problem since the user’s satisfaction function is unknown. Substantive work has been done on this topic where the key idea is to play audio samples to the user, each shaped by a different filter hi, and query the user for their satisfaction scores f(hi). A family of “surrogate” functions is then designed to fit these scores and the optimization method gradually refines these functions to arrive at the filter ˆh* that maximizes satisfaction. In certain applications, we observe that a second type of querying is possible where users can tell us the individual elements h*[j] of the optimal filter h*. Consider an analogy from cooking where the goal is to cook a recipe that maximizes user satisfaction. A user can be asked to score various cooked recipes (e.g., tofu fried rice) or to score individual ingredients (say, salt, sugar, rice, chicken, etc.). Given a budget of B queries, where a query can be of either type, our goal is to find the recipe that will maximize this user’s satisfaction. Our proposal builds on Sparse Gaussian Process Regression (GPR) and shows how a hybrid approach can outperform any one type of querying. Our results are validated through simulations and real world experiments, where volunteers gave feedback on music/speech audio and were able to achieve high satisfaction levels. We believe this idea of hybrid querying opens new problems in black-box optimization and solutions can benefit other applications beyond audio personalization.

The Internet of Things (IoT) has emerged as a key networking infrastructure that connects a large number of sensors, thereby allowing the collection and processing of large amounts of sensor data efficiently. Edge computing involves the deployment of computing devices close to sensor locations in order to process sensor data and derive useful information quickly and efficiently near the source. This improves the quality of the collected data and reduces the strain on the underlying communication networks. The obtained data can be used for a number of inference tasks using machine learning. The majority of machine learning research efforts assume that the collected data is untainted by various effects in the data collection or transmission process. The data-driven models generated with this assumption and the inferences drawn are frequently unusable in practical scenarios. Predictive maintenance is the process of utilizing sensor data obtained from equipment to determine and predict the optimal time for maintenance of the equipment. Using predictive maintenance, the time and money spent on maintaining equipment can be greatly reduced. In this paper, we perform distributed modeling using edge computing to determine the Remaining Useful Lifetime (RUL) of machines used in a manufacturing plant. We analyze the effects of wireless network degradation on the accuracy of the distributed models and propose active learning algorithms to mitigate these effects, thereby improving the usefulness and robustness of predictive maintenance in realistic settings.