Artificial Intelligence and Robotics Commons^™

Extracting Dnn Architectures Via Runtime Profiling On Mobile Gpus, Dong Hyub Kim

Masters Theses

Due to significant investment, research, and development efforts over the past decade, deep neural networks (DNNs) have achieved notable advancements in classification and regression domains. As a result, DNNs are considered valuable intellectual property for artificial intelligence providers. Prior work has demonstrated highly effective model extraction attacks which steal a DNN, dismantling the provider’s business model and paving the way for unethical or malicious activities, such as misuse of personal data, safety risks in critical systems, or spreading misinformation. This thesis explores the feasibility of model extraction attacks on mobile devices using aggregated runtime profiles as a side-channel to leak …

Evaluation Of Robust Deep Learning Pipelines Targeting Low Swap Edge Deployment, David Carter Cornett

Masters Theses

The deep learning technique of convolutional neural networks (CNNs) has greatly advanced the state-of-the-art for computer vision tasks such as image classification and object detection. These solutions rely on large systems leveraging wattage-hungry GPUs to provide the computational power to achieve such performance. However, the size, weight and power (SWaP) requirements of these conventional GPU-based deep learning systems are not suitable when a solution requires deployment to so called "Edge" environments such as autonomous vehicles, unmanned aerial vehicles (UAVs) and smart security cameras.

The objective of this work is to benchmark FPGA-based alternatives to conventional GPU systems that have the …

Neural Network Supervised And Reinforcement Learning For Neurological, Diagnostic, And Modeling Problems, Donald Wunsch Iii

Masters Theses

“As the medical world becomes increasingly intertwined with the tech sphere, machine learning on medical datasets and mathematical models becomes an attractive application. This research looks at the predictive capabilities of neural networks and other machine learning algorithms, and assesses the validity of several feature selection strategies to reduce the negative effects of high dataset dimensionality. Our results indicate that several feature selection methods can maintain high validation and test accuracy on classification tasks, with neural networks performing best, for both single class and multi-class classification applications. This research also evaluates a proof-of-concept application of a deep-Q-learning network (DQN) to …

Controlled Switching In Kalman Filtering And Iterative Learning Controls, He Li

Masters Theses

“Switching is not an uncommon phenomenon in practical systems and processes, for examples, power switches opening and closing, transmissions lifting from low gear to high gear, and air planes crossing different layers in air. Switching can be a disaster to a system since frequent switching between two asymptotically stable subsystems may result in unstable dynamics. On the contrary, switching can be a benefit to a system since controlled switching is sometimes imposed by the designers to achieve desired performance. This encourages the study of system dynamics and performance when undesired switching occurs or controlled switching is imposed. In this research, …

Explorations Into Machine Learning Techniques For Precipitation Nowcasting, Aditya Nagarajan

Masters Theses

Recent advances in cloud-based big-data technologies now makes data driven solutions feasible for increasing numbers of scientific computing applications. One such data driven solution approach is machine learning where patterns in large data sets are brought to the surface by finding complex mathematical relationships within the data. Nowcasting or short-term prediction of rainfall in a given region is an important problem in meteorology. In this thesis we explore the nowcasting problem through a data driven approach by formulating it as a machine learning problem.

State-of-the-art nowcasting systems today are based on numerical models which describe the physical processes leading to …

Fuzzy Adaptive Resonance Theory: Applications And Extensions, Clayton Parker Smith

Masters Theses

"Adaptive Resonance Theory, ART, is a powerful clustering tool for learning arbitrary patterns in a self-organizing manner. In this research, two papers are presented that examine the extensibility and applications of ART. The first paper examines a means to boost ART performance by assigning each cluster a vigilance value, instead of a single value for the whole ART module. A Particle Swarm Optimization technique is used to search for desirable vigilance values. In the second paper, it is shown how ART, and clustering in general, can be a useful tool in preprocessing time series data. Clustering quantization attempts to meaningfully …

Real-Time Mobile Stereo Vision, Bryan Hale Bodkin

Masters Theses

Computer stereo vision is used extract depth information from two aligned cameras and there are a number of hardware and software solutions to solve the stereo correspondence problem. However few solutions are available for inexpensive mobile platforms where power and hardware are major limitations. This Thesis will proposes a method that competes with an existing OpenCV stereo correspondence method in speed and quality, and is able to run on generic multi core CPU’s.