Sparsely Activated BNNs from Deep Gaussian Process
We developed a sparse expansion of deep Gaussian processes as Bayesian neural networks that are amenable to efficient training and prior design.
Weakly Private Information Retrieval
We study the problem of weakly private information retrieval (PIR) when there is heterogeneity in servers’ trustfulness under the maximal leakage (Max-L) metric and mutual information (MI) metric.
AI-driven Dyanmic mmWave Networking
We develop a general framework for effectively deploying reinforcement learning (RL) to control 5G mmWave IAB networks.
Machine Learning and Data Science
We studied the general area of machine learning and data science. A machine learning-based algorithm to is proposed to classify customer reviews. A variant of GAN, WGAN-GP, is designed to extend data set of EEG signals.
Matrix Optimization and Convex Programming
We studied matrix optimization problems using convex programming and machine learning techniques. A complex-valued gradient neural network (CVGNN) is proposed to solve the Moore-Penrose inverse of complex matrices. We also implemented blind deconvolution using convex programming.