Machine learning has enabled significant advancements in diverse fields, yet, with a few exceptions, has had limited impact on computer architecture. Recent work, however, has begun to explore broader application to design, optimization, and simulation. Notably, machine-learning-based strategies often surpass prior state-of-the-art analytical, heuristic, and human-expert approaches. This thesis first...
Machine learning applied to computer architecture has rapidly transitioned from a theoretical novelty to being a driving force behind design, control, and simulation in practically all components. These machine-learning-based methodologies are further notable for their scalability to increasingly complex design challenges, which has allowed these methodologies to surpass the prior...
Various natural language processing (NLP) tasks necessitate deep models that are fast, efficient, and small based on their ultimate application at the edge or elsewhere. While significant investigation has furthered the efficiency and reduced the size of these models, reducing their downstream latency without significant trade-offs remains a difficult task....
The machine learning and deep learning models have been very lightly explored in analyzing the behavior of On-Chip network traffic. These models have proven their potential in pattern recognition, classification etc... In this paper we analyze the spatial pattern that each workload exhibits in its life cycle during execution. We...
Simultaneous speech-to-text translation remains a difficult yet important problem for modern machine learning models whereby a text translation is generated concurrently with receiving partial speech inputs. One state-of-the-art simultaneous speech-to-text model is the augmented memory transformer whose encoder breaks a speech input into fixed-size overlapping segments composed of left, right,...
In recent years, RF (Radio Frequency) device fingerprinting using deep learning has emerged as a method of identifying devices solely by their RF transmissions. Conventional approaches to this type of device fingerprinting are not portable to different domains. That is, if a model for this purpose is trained on data...
Optical wireless communication (OWC) is an alternative to radio frequency (RF)communication with a signi cantly larger and unregulated spectrum. In OWC systems, optical orthogonal frequency division multiplexing (O-OFDM) with intensity modulation and direct detection (IM/DD) is commonly used. There are two common signal structures in most OWC systems based on...
Extensive studies have been undertaken on the transient stability of large interconnected
power systems with flexible ac transmission systems (FACTS) devices installed.
Varieties of control methodologies have been proposed to stabilize the postfault system
which would otherwise eventually lose stability without a proper control. Generally speaking,
regular transient stability is...
An adaptive pitch axis autopilot design procedure is presented. The
design procedure is applicable to both stable and unstable pitch axis models
and to those having nonminimum phase. The design approach assumes
the adaptive autopilot is activated after achieving level flight. It is shown a
rate-feedback compensator can be designed...
The harmonic problem in power systems is gaining more attention as incidences correlated with harmonics increase. Conventional passive filtering techniques for harmonic mitigation have inherent problems, and purely active filters have the disadvantages of higher costs and ratings. Hybrid active filters inherit the efficiency of passive filters and the improved...