Engineering Commons^™

Memristive Architectures And Algorithms For Approximate Graph-Based Inference, Mohammad M.A. Taha

Dissertations and Theses

The goal of this thesis is to design fast, low-power, robust graph-based inference systems. Our approach to this is by using (1) in-memory computing, (2) approximate computing principles, and (3) memristive devices.

This work is motivated by the fact that conventional von Neumann architectures are not efficient for inference applications, mainly due to the data transfer bottleneck. Adding cache memories and using GPUs is a remedy, however, does not eliminate this bottleneck. In-memory computing is an alternative approach that performs all computations inside the memory, thus eliminating the data transfer bottleneck.

The memristor, which is a passive two-terminal device, is …

Absorptive Reflectionless Filters, Guy Barry Lemire

Dissertations and Theses

For many applications requiring some sort of signal filtering or signal conditioning, the filter requirements are usually approached with a single purpose in mind, which is to maximize both passband signal amplitude and stop band signal attenuation to the load with little to no thought given to what happens to the stop band signal energy. Many conventional filters have very poor impedance matching in the stopband resulting in reflected energy or large return loss (S₁₁). This reflected energy can then cause interactions with adjoining system components which do not in general respond well to spurious reflected energy and …