Lab Director
Guni Sharon
Assistant professor, Texas A&M University, Department of Computer Science & Engineering
Research Interests: Artificial Intelligence, Intelligent transportation systems, Reinforcement learning, Combinatorial optimization
Website: [ http://faculty.cse.tamu.edu/guni/ ]
Dr. Sharon is a researcher with a strong theoretical basis in artificial
intelligence. Specifically, reinforcement learning, combinatorial
search, multiagent route assignment, game theory, flow and convex
optimization, and multiagent modeling and simulation. He gained vast
knowledge and experience in utilizing his theoretical foundations
towards traffic management and traffic optimization application.
Dr. Sharon strives to further the impact of his applicable
expertise for solving real-life problems while simultaneously
continuing to make theoretical advances that justify the proposed
solutions.
PhD Students
James Ault (entered Fall 2019)
Research Interests: Reinforcement learning, Multiagent Systems, Human-agent Interaction
Website: [ http://people.tamu.edu/~jault/ ]
James is interested in solving the key challenges preventing the widespread application of
reinforcement learning techniques to real-world problems. Currently, he is working towards
learning interpretable policies for traffic signal controllers to reduce the delay caused
by intersections. He earned a bachelor's degree in Computer Science from Rutgers University
and began his graduate education at Texas A&M. Outside of research, he enjoys hiking and
competitive strategy games.
- Cooperative Multi-agent Reinforcement Learning Applied to Multi-intersection Traffic Signal Control (Workshop paper).
James Ault and Guni Sharon
In The Third Workshop on Data-driven Intelligent Transportation at CIKM 2022, 2022
pdf bib - A Framework for Predictable Actor-Critic Control (Workshop paper).
Josiah Coad, James Ault, Jeff Hykin, and Guni Sharon
In Proceedings of the Deep Reinforcement Learning (DRL) Workshop at NeurIPS 2022, 2022
pdf bib - Reinforcement Learning Benchmarks for Traffic Signal Control.
James Ault and Guni Sharon
In Proceedings of the 35th Neural Information Processing Systems (NeurIPS 2021) Track on Datasets and Benchmarks, 2021
pdf bib - Agent-Based Markov Modeling for Improved COVID-19 Mitigation Policies.
Roberto Capobianco and Varun Kompella and James Ault and Guni Sharon and Stacy Jong and Spencer Fox and Lauren Meyers and Peter R. Wurman and Peter Stone
In Journal of Artificial Intelligence Research (JAIR), 2021
pdf bib code - Multiagent Epidemiologic Inference through Realtime Contact Tracing.
Guni Sharon, James Ault, Peter Stone, Varun Kompella, and Roberto Capobianco
In the Proceedings of the 20th International Conference on Autonomous Agents and MultiAgent Systems (AAMAS 2021), 2021
pdf bib code - Learning an Interpretable Traffic Signal Control Policy.
James Ault, Josiah P. Hanna, and Guni Sharon
In the Proceedings of the 19th International Conference on Autonomous Agents and MultiAgent Systems (AAMAS 2020), 2020
pdf bib code
Sheelabhadra "Sheel" Dey (entered Fall 2019)
Research Interests: Reinforcement learning, Computer vision, Robotics
Website: [ https://sheelabhadra.github.io/ ]
Sheelabhadra "Sheel" Dey is interested in
building agents that can perceive their surroundings, quickly learn
to navigate efficiently, and then act safely. He is currently
working on developing end to end reinforcement learning algorithms
for autonomous driving. He has a background in computer vision and
is interested in the combined framework of perception and planning.
He holds a master's degree in computer science from Texas A&M and a
bachelor's degree in electronics and communication engineering from
National Institute of Technology, Trichy, India. Outside of
research, he enjoys playing the guitar and sketching.
Sumedh Pendurkar
(entered Fall 2020)
Research Interests: Reinforcement Learning, Combinatorial
Optimization, Heuristic Search
Website: [ https://sumedhpendurkar.github.io/ ]
Sumedh is interested to leverage advances in machine learning to
enhance the performance of algorithms used for combinatorial
optimization problems. Currently, his work focuses on learning the
heuristic function approximators for various heuristic search
algorithms that result in faster learning while maintaining
optimal search guarantees. Outside of research, he enjoys
playing guitar and competitive video games.
- The (Un)Scalability of Informed Heuristic Function
Estimation in NP-Hard Search Problems
Sumedh Pendurkar, Taoan Huang, Brendan Juba, Jiapeng Zhang, Sven Koenig, and Guni Sharon
In Transactions on Machine Learning Research, 2023
pdf bib code Talk - Comparison between popular genetic algorithm
(GA)-based tool and covariance matrix adaptation –
evolutionary strategy (CMA-ES) for optimizing indoor daylight
Manal Anis, Sumedh Pendurkar, Yun Kyu Yi, and Guni Sharon
In Proceedings of Building Simulation 2023: 18th Conference of IBPSA., 2023
pdf bib code - Bilevel Entropy based Mechanism Design for Balancing Meta in Video Games.
Sumedh Pendurkar, Chris Chow, Luo Jie, and Guni Sharon
In Proceedings of the 22th International Conference on Autonomous Agents and MultiAgent Systems (AAMAS), 2023
pdf bib code - The (Un)Scalability of Heuristic Approximators for NP-Hard Search Problems (Workshop paper).
Sumedh Pendurkar, Taoan Huang, Sven Koenig, and Guni Sharon
In Proceedings of the I (Still) Can't Believe It's Not Better! Workshop at NeurIPS 2022, 2022
pdf bib - A Discussion on the Scalability of Heuristic Approximators.
Sumedh Pendurkar, Taoan Huang, Sven Koenig, and Guni Sharon
In Proceedings of the International Symposium on Combinatorial Search, 2022
pdf bib - A Joint Imitation-Reinforcement Learning Framework for Reduced Baseline Regret.
Sheelabhadra Dey, Sumedh Pendurkar, Guni Sharon, and Josiah Hanna
In the Proceedings of the 34th International Conference on Intelligent Robots and Systems (IROS 2021), 2021
pdf bib slides demo code
Vaibhav Bajaj (entered Summer 2021)
Research Interests: Reinforcement Learning, Multiagent Systems, Behaviour-based Learning
Website: [
Vaibhav is interested in developing techniques that make it possible
for an agent to generalize its learning strategy and performance
across multiple tasks and environments while coordinating with other
agents in the environment. Currently, he is working towards
utilizing Curriculum and Knowledge transfer based techniques to
allow autonomous agents to learn to perform complex tasks easily
and perform well with different team-sizes and environments. He has
a background in multi-agent obstacle avoidance and path planning and
is interested in learning behaviours that make it possible for
autonomous agents to easily mix into everyday Human life. He
received his Bachelors in Computer Science and Engineering from PES
University, Bangalore, India. Outside of research, he enjoys playing
Badminton and Cooking.
Jeff Hykin (entered Summer 2021)
Research Interests: Reinforcement Learning, Spiking Networks, Learning without Backpropagation
Website: [https://jeffhykin.wixsite.com]
Jeff is interested in fully solving robotic spatial awareness. However, rather than accuracy, his focus is on naturally robust learning techniques: methods that are effectively immune to adversarial attacks and result in a high-level understanding of the environment. His work involves multi-task learning, curriculum learning, knowledge transfer, state-space compression, and virtual environments. He enjoys frequently publishing small packages to continually accelerate and simplify development.
His first few years of undergraduate were spent studying cognitive science, psychology, and neuroscience at the University of Texas at Dallas. Later he transferred to Texas A&M University, completed his bachelor's in Computer Science, and is now continuing at A&M for his Ph.D. in Computer Science.
With over 100 GitHub repositories, Jeff likes to do open source work on everything from VS Code extensions to graph-based database interfaces. Outside of programming, he enjoys rock climbing, sand volleyball, soccer, skydiving, and reading.