Robust and Adaptive Deep Learning via Bayesian Principles

Enhancement of Robust Tracking Performance via Switching Supervisory Adaptive Control

When the process is highly uncertain, even linear minimum phase systems must sacrifice desirable feedback control benefits to avoid an excessive ‘cost of feedback’, while preserving the robust stability. In this paper, the problem of supervisory based switching Quantitative Feedback Theory (QFT) control is proposed for the control of highly uncertain plants. According to this strategy, the unce...

Bayesian Deep Learning | Uncertainty in Deep Learning

Bayesian Deep Q-Learning via Continuous-Time Flows

Efficient exploration in reinforcement learning (RL) can be achieved by incorporating uncertainty into model predictions. Bayesian deep Q-learning provides a principle way for this by modeling Q-values as probability distributions. We propose an efficient algorithm for Bayesian deep Q-learning by posterior sampling actions in the Q-function via continuous-time flows (CTFs), achieving efficient ...

Robust Face Recognition via Block Sparse Bayesian Learning

Face recognition (FR) is an important task in pattern recognition and computer vision. Sparse representation (SR) has been demonstrated to be a powerful framework for FR. In general, an SR algorithm treats each face in a training dataset as a basis function, and tries to find a sparse representation of a test face under these basis functions. The sparse representation coefficients then provide ...

Adaptive Static Analysis via Learning with Bayesian Optimization

ions are binary vectors with indices in JP , and are ordered pointwise: a v a′ ⇐⇒ ∀j ∈ JP . aj ≤ aj . Intuitively, JP consists of the parts of P where we have switches for controlling the precision of an analysis. For instance, in a partially context-sensitive analysis, JP is the set of procedures or call sites in the program. In our partially ƒow-sensitive analysis, it denotes the set of progr...