Prithviraj (Raj) Dasgupta
Computer Engineer | Distributed Intelligent Systems Section | Naval Research Laboratory, Washington D. C.
I am a research scientist in the Distributed Intelligent Systems Section, Information Technology Division at the Naval Research Laboratory in Washington, D. C.
My group does research in the field of AI and machine learning around the areas of adversarial AI, reinforcement learning, game theory and multi-robot/multi-agent systems.
From 2001-2019, I was the Union Pacific Endowed Professor (tenured) with the Computer Science Department at the University of Nebraska, Omaha. I had established and directed the CMANTIC Robotics Lab there and led several large projects supported by NASA, Office of Naval Research and NAVAIR . I received the highest research award called ADROCA at the university in 2017. I also served as the Director of the Ph.D. in Information Technology program at the university from 2010-2014. I am a senior member of IEEE.
My full list of publications from Google Scholar Profile and DBLP
Recent News
March 2023: Journal paper in International Journal of Serious Games: P. Dasgupta and J. Kliem, "Improved Reinforcement Learning in Asymmetric Real-time Strategy Games via Strategy Diversity: A Case Study for Hunting-of-the-Plark Game, " International Journal of Serious Games, vol, 10(1), pp. 19--38. [Link to paper]
January 2023: Our paper titled "Generating synthetic data sequences for human-AI interactions-based tasks via reward shaping" was accepted for oral presentation at Synthetic Data Generation for AI/ML Conference, SPIE Defense + Commercial Sensing 2023. [Paper]
November 2022: I was an invited panelist at the Workshop on Adversarial Machine Learning and Cybersecurity at MILCOM 2022.
October 2022: Our team did successful physical experiments with USVs controlled by our reinforcement learning algorithms for playing a marine Capture-the-Flag game, at USMA, West Point.
April 2022: We participated with a booth on Naval Applications of AI at the STEM Expo at the Sea/Air/Space Exposition in National Harbor, MD. It was great to see the enthusiasm from the community and K-12 students on our research.
November 2021: I was invited to give a talk on Robust Adversarial AI at DoD C4I Colloquium.
October 2021: Two papers/presentations on our research on Robust Adversarial Imitation Learning and Opposed AI at the NATO Symposium on Artificial Intelligence, Machine Learning and Big Data for Hybrid Military Operations (AI4HMO); one presentation on our research on adversarial malware generation at the NATO Symposium on Cybersecurity
April 2021: Paper on Adversarial Imitation Learning using Options at AI/ML for Multi Domain Operations Conference, SPIE Defense + Commercial Sensing, 2021, Link
March 2021: U. S. Patent application "System and Method for Improving Classification in Adversarial Machine Learning." U.S. Patent Application No. 17/188,923. . Details of our approach is described in our paper here.
Recent Publications
B. Brandt and P. Dasgupta, "Generating Synthetic Data sequences for Human-AI Interactions-based Tasks via Reward Shaping," SPIE Defense + Commercial Sensing 2023, Conference on Synthetic Data for Artificial Intelligence and Machine Learning: Tools, Techniques, and Applications. [arxiv link]
P. Dasgupta and J. Kliem, "Improved Reinforcement Learning in Asymmetric Real-time Strategy Games via Strategy Diversity: A Case Study for Hunting-of-the-Plark Game, " International Journal of Serious Games, vol, 10(1), pp. 19--38. [Link to paper]
G. Perrotta, R. Gardner, C. Lowman, M. Taufeeque, N. Tongia, S. Kalyanakrishnan, G. Clark, K. Wang, E. Rothberg, B. Garrison, P. Dasgupta, C. Canavan, L. McCabe, "The Second NeurIPS Tournament of Reconnaissance Blind Chess," Proceedings of the NeurIPS 2021 Competitions and Demonstrations Track, PMLR 176:53-65, 2022 . [Link]
P. Spencer, P. Dasgupta, M. McCarrick, M. Novitzky, D. Hubczenko, A. Jeffery, S. Redfield, J. James, R. Mittu, "Opposed Artificial Intelligence: Developing Robustness to Adversarial Attacks in Attacker-Defender Games via AI-based Strategic Game-Playing ," Proc. NATO STO Symposium on AI, Machine Learning and Big Data for Hybrid Military Operations, 2021.
P. Dasgupta, "Playing Games to Learn Robustly from Adversarial Expert Demonstrations," Proc. NATO STO Symposium on AI, Machine Learning and Big Data for Hybrid Military Operations, 2021.
P. Dasgupta, "Using options to improve robustness of imitation learning against adversarial attacks," Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications III. Vol. 11746. SPIE, 2021. [Link]
P. Dasgupta, and Z. Osman. "A Comparison of State-of-the-Art Techniques for Generating Adversarial Malware Binaries." arXiv preprint arXiv:2111.11487 (2021). [Link]
J, B, Collins and P. Dasgupta. "Rethinking Intelligent Behavior as Competitive Games for Handling Adversarial Challenges to Machine Learning." Adversary-Aware Learning Techniques and Trends in Cybersecurity. Springer, Cham, 2020. 3-16. [Link]
P. Dasgupta, J. B. Collins, and M. McCarrick. "Improving costs and robustness of machine learning classifiers against adversarial attacks via self play of repeated Bayesian games." The Thirty-Third International FLAIRS Conference. 2020. [Link]
P. Dasgupta, and J. B. Collins. "A survey of game theoretic approaches for adversarial machine learning in cybersecurity tasks." AI Magazine 40.2 (2019): 31-43. [Link]
Email: pdasgupta-at-ieee-dot-org