Welcome!

I am Shivvrat Arya, a Ph.D. graduate in Computer Science from The University of Texas at Dallas. My research lies at the intersection of neurosymbolic AI, probabilistic reasoning, and computer vision. I collaborate with Dr. Vibhav Gogate and Dr. Yu Xiang on several DARPA-funded projects:

• DARPA Explainable Artificial Intelligence (XAI)
• DARPA Perceptually-Enabled Task Guidance (PTG)
• DARPA Assured Neuro Symbolic Learning and Reasoning (ANSR)

Research Focus

Broadly speaking, my research interests primarily focus on artificial intelligence and machine learning, with an emphasis on neuro-symbolic and graph-based methods. I aim to develop AI systems that combine interpretability with powerful reasoning capabilities. Key areas include:

• Neurosymbolic AI: Integrating formal world models with neural networks
• Probabilistic Methods: Novel inference approaches for graphical models
• Explainable AI: Interpretable models supporting polynomial-time reasoning
• Combinatorial Optimization: Efficient algorithms for NP-hard problems
• Computer Vision: Advanced video understanding and activity recognition

Highlights

• Published research recognized with best paper awards, spotlight, and oral presentations
• Developed novel neural network-based inference methods for probabilistic models
• Released NeuPI, a neural network-based inference engine for probabilistic models
• Built real-time AR guidance systems for complex physical tasks
• Created CaptainCook4D, an egocentric 4D dataset for procedural task understanding

Education

• Ph.D., Computer Science – The University of Texas at Dallas
• M.S., Computer Science – The University of Texas at Dallas
• B.Tech., Computer Science and Engineering – IIIT Vadodara