CS&E Ph.D. students named Doctoral Dissertation Fellows

Six Ph.D. students working with CS&E professors have been named Doctoral Dissertation Fellows for the 2019-2020 school year. The Doctoral Dissertation Fellowship is a highly competitive fellowship that gives the University’s most accomplished Ph.D. candidates an opportunity to devote full-time effort to an outstanding research project by providing time to finalize and write a dissertation during the fellowship year. 

The award includes a stipend of $25,000, tuition for up to 14 thesis credits each semester, and subsidized health insurance through the Graduate Assistant Health Plan. More details on the award and nomination process can found on the Graduate School’s website.

CS&E congratulates the following students on this outstanding accomplishment:

Karan Aggarwal
Advisor: Jaideep Srivastava
Models for Mining Human Activity and Sleep Patterns with Applications to Diagnosis and Monitoring
Aggarwal’s research aims to exploit the power of machine learning models for management of chronic conditions, particularly sleep related. His research seeks to transition from the “reactive” traditional health-care approach to a more “pro-active” health-care delivery by harnessing the unprecedented access to the activity and sleep data with connected medical devices.

Wenbo Dong
Advisor: Volkan Isler
A Robotic Vision Approach to Next-Generation Automated Precision Agriculture
Vision sensors mounted on mobile robotic platforms hold great promise in automated agriculture management. However, established computer vision techniques employed to acquire and process digital images fail to perform well in agricultural environments due to the environmental complexity. Dong has designed and developed 3D computer vision algorithms that improve the accuracy of image acquisition, suppress undesirable environmental interferences, and generate accurate and precise 3D models of plants with detailed information automatically extracted. He hopes his research will help propel automated agricultural practices by building real-time 3D “Google Maps” of farmlands, crops fields, and orchards.

Md Jahidul Islam
Advisor: Junaed Sattar
Perceiving Human Motions, Gestures, and Actions for Human-Robot Collaboration
Islam designs visual perception techniques in order for autonomous robots to efficiently and naturally interact with humans. The perception capabilities include the detection and interpretation of human motions, poses, gestures, and activities in social settings and in unstructured environments, such as underwater. These are challenging problems to both academic and industrial researchers, particularly for embedded platforms under real-time operating constraints. To this end, his research at the Interactive Robotics and Vision Laboratory aims at finding or improving solutions to these problems by leveraging computer vision, machine learning, and deep learning-based techniques.

Xiaowei Jia
Advisor: Vipin Kumar
Physics-guided Machine Learning for Scientific Knowledge Discovery
Physics-based models are often used to study engineered and natural systems. Despite their extensive use, these models have well-known limitations due to incomplete or inaccurate representations of the physical processes being modeled. Machine learning (ML) methods are being increasingly considered as an alternative but also have limitations, e.g., they only capture statistical relationships from limited data and thus can produce physically-inconsistent results. Jia’s dissertation focuses on integrating physics into ML models with the aim of significantly improving the prediction and generalization ability using limited data while ensuring physical consistency.

Ibrahim Sabek
Advisor: Mohamed Mokbel
Adopting Markov Logic Networks for Big Spatial Data and Applications
Spatial data has become ubiquitous, e.g., GPS data, medical data, with increasingly sheer sizes. This raises the need for efficient spatial analysis solutions to extract useful insights from such data. Meanwhile, Markov Logic Networks (MLN) have emerged as powerful framework for building usable and scalable machine learning tools. Unfortunately, MLN is ill-equipped for spatial applications because it ignores the distinguished spatial data characteristics. Sabek’s research seeks a way to provide native support for spatial data inside MLN and help build scalable spatial analysis and knowledge construction tools.

Vaibhav Sharma
Advisor: Stephen McCamant
Using Efficient Symbolic Execution To Find Substitutable Code
Bugs in commercial software and third-party components are an undesirable and expensive phenomenon. Such software is usually released to users only in binary form. The lack of source code renders users of such software dependent on their software vendors to repair bugs. Such dependence is even more harmful if the bugs introduce new vulnerabilities in the software. Being able to automatically repair security and functionality bugs breaks this dependence and increases software robustness. Sharma’s research seeks to develop a binary program repair tool that uses existing bug-free fragments of code to repair buggy code.