A Platform for AI and Deep Learning
Abstract
St. MaryÕs University is a federally designated Hispanic Serving Institution (HSI) located in San Antonio, Texas. The Engineering Department at St. MaryÕs aims to develop an Artificial Intelligence (AI) and deep learning platform that will significantly enhance the UniversityÕs capacity for research and technical training. The requested equipment includes: AMAX ServMax Xeon Server with 24 Tesla V100 16GB HBM2; AMAX Cluster; Data Center Manager Appliance; and power distribution unit (PDU). Developing effective instrument-based education infrastructure will impact not only several hundred St. MaryÕs students annually (predominantly from groups underrepresented in STEM fields), but also a broader audience from other U.S. universities or institutions as students will be able to securely access the platform over the internet to run AI and deep learning simulations off campus. At least 10 faculty members across the departments of Engineering, Computer Science, Physics, and Math will be able to incorporate the equipment into their research and/or teaching. Providing research opportunities to undergraduates is a priority at St. MaryÕs. Developing the proposed platform will expose students to cutting-edge research in AI and deep learning, which will help to prepare them for graduate school and/or industry. Of the many potential research areas that could be explored using this equipment, examples include application of AI and deep learning in pattern recognition, data analysis, and cybersecurity. The PIÕs research involves various aspects of deep learning in pattern recognition and data analysis. His research is aimed at developing, evaluating, and applying deep neural networks in classification, clustering, human behavior, outlier detection, and optimization. Furthermore, the proposed platform expands the PI and other facultyÕs current research in fields such as health monitoring, image processing, robotics, signal and image processing, pattern analysis, data analysis, data mining, optimization, and information security. The proposed platform will exponentially increase our processing power, thus enabling us to run more complex simulations. In cybersecurity, it is known that a large majority of the new threats and malware are a small mutation of existing ones; however, current cybersecurity solutions have difficulty in detecting a significant percentage of the new malware. This results in vulnerabilities that leave institutions/organizations exposed to information breaches, data corruption, data theft, and cyber deception. St. MaryÕs plans to address these problems by applying deep learning techniques to cybersecurity. These techniques are used to detect malware and to classify a userÕs behavior as legitimate or malicious according to known features used to train the deep neural network that resembles or mimics brain structure. The equipment acquired through this grant will be used to train the deep neural networks to prevent cyberattack without any human intervention. The proposed AI and deep learning platform will enhance the quality of science and engineering programs in the following ways: (1) create an infrastructure for deep learning research and teaching; (2) increase research capacity, productivity and research partnership opportunities; (3) introduce students to interdisciplinary research and education; (4) introduce new techniques to pedagogical methods of teaching machine and deep learning; (5) develop new proficiencies for faculty and students in the use of current cutting-edge technology; (6) provide an exciting science perspective to potentially attract high school students to careers in science and engineering; and (7) improve skills in problem solving, multidisciplinary teamwork, and application of math, science, and engineering principles.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 21, 2019
- Source ID
- W911NF1910159
Entities
People
- Vahid Emamian
Organizations
- Army Contracting Command
- United States Army