First Year Projects
Anahita Khojandi – Department of Industrial & Systems Engineering
Dynamic Deep Reinforcement Learning-Bayesian Framework
The modern economy has been described as sensor-based. Sensors can collect data at high-frequency, which could potentially allow decision making in real time. Many frameworks exist that allow for such decision making, but they are typically either really good at fusing many data streams or adapting to changes in a dynamic environment. Khojandi’s work proposes to integrate the strengths of these frameworks and develop new methods that can accommodate various human-in-the-loop considerations and resource limitations.
Michela Taufer – Min H. Kao Department of Electrical Engineering and Computer Science
Empowering Training and Validation Stages in AI-Orchestrated Workflows
Adversarial attacks on AI-orchestrated systems occur when small perturbations of an input cause a dramatic change in the prediction of a neural network.There is recent research around training neural networks to be resilient to attacks, but none of these methods can prevent adversarial attacks. Taufer proposes to research methods to detect adversarial attacks so that the attack can be mitigated.
Second Year Projects
Barry Bruce – Department of Biochemistry & Cellular and Molecular Biology
Probing Dynamics of Photosystem | Oligomerization Using SANS and Cryo-EM
On earth, the process of photosynthesis has had to evolve highly adaptive processes to deal with light levels that regularly fluctuate. These organisms are able to respond to changing environmental stimuli to optimize growth in highly variable conditions. Bruce’s project aims to develop an imaging platform for investigating the multiscale phenomena associated with the photoconversion process.
Jamie Coble – Department of Nuclear Engineering
Development of Capacitive Dimensional Change Sensors for Nuclear Materials Measurement
There is a critical need for sensors that provide real-time data regarding material evolution under high accelerated irradiation. One such sensor is currently under development at ORNL. Coble’s project proposes to create a complementary sensor to provide an alternate measurement modality for in-pile measurement of dimensional change.
Mark Dean – Min H. Kao Department of Electrical Engineering and Computer Science
Next-generation Neuromorphic Coprocessor Power Consumption in the Beyond Exascale Era
The von Neumann computer architecture has been the basis for computer design for more than half a century, but in recent years neuromorphic computing has emerged as a compliment to that architecture. The primary drawback of these systems currently is the communication between elements and processing units. Dean proposes to develop a new configuration of these elements and processors to address this issue.
Seddik Djouadi – Min H. Kao Department of Electrical Engineering and Computer Science
Robust Control Design for Power Electronics-Enabled Grid Architectures
Modern power networks are made up of classical electromechanical machines and renewable energy resources interconnected through electronic devices, resulting in characteristics that inhibit performance. Djouadi proposes the creation of a framework to maximize controllability and ensure safe system operations.
Claudia Rawn – Department of Materials Science and Engineering
Complementary X-ray Diffraction Studies for the Characterization of Chromium Dissolution into Molten Chloride Salts
Rawn’s project seeks to leverage the Diffraction Facility at the Joint Institute of Advanced Materials to study the effects of high heat on chromium and molten salts under high heat conditions. The data uncovered may help advance the understanding of the fundamental science behind the depletion of chromium in alloys.
Sarah Werner – Department of Microbiology
Mining GWAVA for Key Factors Shaping Microbiome Structure
Recent studies have demonstrated a clear overlap in the internal microbial communities among plants, the interactions of which contribute to root microbiome composition. Werner’s project aims to identify and test predictions to define a new role for the bacteria Streptomycetacea in the root microbiome.
Steven Wilhelm – Department of Microbiology
Microeukaryotes and Their Viruses : Uncovering Their Hidden Role in One of the Largest Terrestrial Carbon Sinks
In the last two decades, it has become clear that microbes drive nearly all of the major biogeochemical processes on the planet, including activities within microbiomes that range from individual multicellular organisms to entire ecosystems. Wilhelm aims to statistically determine which viruses infect which cells in a microbiome, outside of lab culture efforts, and extend that information to functional processes.