Building an AI Laboratory at GSU for Research, Education and Outreach
Abstract
This proposal requests for the acquisition of an A100 GPU cluster, a NeXus-32 multimodal biosensing system, a Clearpath Ridgeback robot platform with a UR5e arm, and ten Xilinx FPGA evaluation kits for establishing the first Artificial Intelligence (AI) laboratory at Georgia State University (GSU), a Title III-V Minority-Serving Institution (MSI). The laboratory will provide both large-scale GPU computing capabilities as well as experimental platforms for the design, development, and demonstration of AI systems in a wide range of areas from robotics to material design. Although GSU is one of the top five largest universities in the nation by enrollment (nearly 54,000 student population), GSU currently does not have a dedicated AI laboratory. The acquisition of the requested equipment will not only mark the first state-of-the-art AI facility at GSU but will also offer high impactful multidisciplinary research activities in the fields including but not limited to brain-computer interface (BCI), human-robot collaboration, real-time embedded systems, drug discovery, material design, geosciences, healthcare, and computational epidemiology. AI is of paramount importance for U.S. economy, national security, and our quality of life. Developing AI systems that automate decision-making in dynamic environments is the focus of PIÕs research group. Currently, our research group has developed an array of deep learning algorithms in the areas of computer vision (CV) and natural language processing (NLP). These algorithms learn from a variety of data sources (e.g., image, video, audio, and text), and span various AI research areas, including explainable AI (XAI), adversarial deep learning, energy-based models, graph neural networks, and model compression. The requested equipment will enable us to enhance and expand our research capacities to brain signal decoding for robot navigation and manipulation, human-robot collaboration via natural language commands and visual demonstration, and deployment of AI to real-time embedded systems. The versatile nature of the requested equipment and broad range of AI applications will also open up multidisciplinary collaborations with various departments, such as Neuroscience, Chemistry, Biology, Geosciences, Physics & Astronomy both within and outside of GSU, for healthcare, drug discovery, material design, epidemic simulation and prediction, etc. The requested equipment will also enhance the capacity of GSU to participate in future DoD research programs and activities. One of the educational goals of the new laboratory is to train both undergraduate and graduate students and enable them to advance their respective fields. Our research work involves statistical modeling, numerical analysis as well as practical programming and experimentation, and thus will provide training opportunities in all three aspects of the research work. Students will learn how to analyze and design deep learning algorithms for a variety of AI applications and systems, as well as gain hands on experience using advanced GPU cluster, BCI, robotics and FPGA platforms. Furthermore, the PI and Co-PIs regularly teach the subjects of AI and project-based courses at GSU. The AI laboratory will not only support various courses including deep learning, reinforcement learning, robotics and embedded systems but also enable the addition of new courses to our curriculum. As such, the requested equipment will significantly improve the quality of the educational program at GSU. Also, the lab will be used for promoting STEM education to both local high school students as well as the large student body of GSU via seminars, workshops and lab visits.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 02, 2022
- Source ID
- W911NF2210196
Entities
People
- Shihao Ji
Organizations
- Army Contracting Command
- Georgia State University
- United States Army