Two Doctoral Students Awarded 2021 Computing Innovation Fellowships
Two doctoral students in Computer Science and Engineering have received the 2021 Computing Innovation CI Fellowships sponsored by the Computing Research Association (CRA) and the Computing Community Consortium (CCC) with the support from the National Science Foundation (NSF).
Applicants' research is evaluated on the intellectual merit and the potential to advance knowledge and the broader impacts to benefit society and contribute to specific societal outcomes.
Ghulam Jilani Quadri
Ghulam's research concentrates on constructing a visualization framework that considers human perceptions for encoding techniques and the task being performed, optimizing visualization design. It would abet designer improved objective guidance with better quality, maximize efficiency, and high decision-making confidence. Users and designers with assorted experiences would be benefited as the diverse participatory design will have visualization effectiveness that could overcome historic inequities in the way designs have been evaluated.
The approach facilitating the proposed framework with mass critical datasets and populations from distinct sources - education and economic domain; combined with visual designs will help designers make better choices, minimize misleading data visualization, and make data communication accessible and efficient. The proposed framework will have a broader impact on data communication engineering and diversifying public engagement in massive decision-making processes. Beyond direct collaborations, any field that deals with data communication could benefit, including data accessibility, accessible communication, engineering fields requiring simulation and visualization---medicine, biomedical, weather forecasting.
“My vision is to extend the current, underlying framework to various cognitive and perceptual measures. I want my research to be more viable and reach a larger audience by concisely communicating critical information from massive data to build situational awareness of significant events. I am grateful for the award and express my gratitude to my advisor Dr. Paul Rosen for his continuous support and guidance. I am also looking forward to working with my postdoc mentor Dr. Danielle Szafir,” said Ghulam.
Ghulam Jilani Quadri is a doctoral candidate in Computer Science and Engineering, advised by Dr. Paul Rosen. Ghulam's area of interest is Information visualization, perceptual visualization, and Human-Computer Interaction. He believes in achieving new avenues and continuously thrives on increasing his horizon. His current research and plan are to provide an optimal and effective data visualization and communication solution. He briefly talks about his accomplishments and accolades on his website.
"I see myself working as a faculty to a research-oriented university and continue my research on exploring visualization optimization to make a perpetual impact on data communication and accessibility," said Ghulam.
His CI mentor is Danielle Albers Szafit at the University of North Carolina at Chapel Hill. His research title is: Developing Perceptual Framework for Task-Optimized Visualization.
Troi's research develops methods for learning state-dependent sensor measurement models (SDSMMs). These models predict the bias and uncertainty in sensor measurements dynamically (for example, the measurement bias and uncertainty for distance-measuring camera sensors). A robot can, in turn, use the predicted bias and noise to determine the accuracy of its sensors, determine optimal trajectories for achieving a goal, or improve the accuracy of a task (such as localization).
The core idea behind this research is that sensor measurement bias and uncertainty are a function of the state of the sensor and its surrounding environment. For example, the measurement uncertainty for a camera sensor can change due to states like the ambient lighting, the camera's speed, and the distance between the camera and the objects it is observing. The goal of the SDSMM framework is to learn how combinations of these states affect measurement bias and uncertainty. Once learned, the model predicts the measurement bias and uncertainty given a state.
"I plan to extend the current framework to other fundamental areas in robotics. For now, one area is planning, where my CIFellow mentor and I intend to combine our frameworks to develop more safe and self-aware robots," said Troi. "I am also interested in improving the current framework's ease of use such that it is more widely adopted. Furthermore, this postdoctoral experience will allow me to build novel lines of research that I can continue to develop as tenure-track faculty at a university."
Since Troi's research focuses on sensors, it impacts the robotics community and the broader global society. For example, robots (such as autonomous vacuums, cars, and drones) rely on sensor measurements to perceive their surroundings and execute tasks. Since many commodity sensors are inherently biased and noisy, his research allows robots to predict the bias and uncertainty dynamically, potentially improving task accuracy and safety.
"After I complete the postdoc, I hope to earn a tenure-track faculty position at a research-oriented university," said Troi. "My goal is to continue exploring state-dependent sensor measurement models, their impact on fundamental areas in robotics, and their application to various types of autonomous vehicles (such as underwater, ground, and aerial vehicles)."
"I am fortunate to have this award and grateful for the opportunity to work at the University of Maryland with Dr. Pratap Tokekar. I am also grateful for the continual support from Dr. Yu Sun, my USF advisor."
His CI Mentor is Pratap Tokekar at the University of Maryland. His research title is: Integrating State-Dependent Sensor Measurement Models and Risk-Aware Planning.