Suguman Bansal
I am an (incoming) Assistant Professor in the School of Computing at Georgia Institute of Techology, starting in Spring 2023. Previously, I was an NSF/CRA Computing Innovation Postdoctoral Fellow at University of Pennsylvania, mentored by Rajeev Alur. I completed Ph.D at Rice University, advised by Moshe Y. Vardi.
I am broadly interested in Formal Methods and Logic, and their applications to Artificial Intelligence, Machine Learning, and Programming Languages. The long-term reserach goal is to build Trustworthy AI by simplifying the design-development-deployment pipeline of safe and reliable AI systems. My current research thrusts include:
I have open positions for interns, graduate students (PhDs and Masters) and postdocs.
Please Apply To GaTech and get in touch if you are interested in
(a) Theory and applications of Formal Methods,
(b) (Deep) Reinforcement Learning, and/or
(c) Automata and Logical Methods.
Check out our Reseach Statement. Please send me your CV, if interested.
Selected Awards and Honors
News
- Tutorial on Specification-Guided Reinforcement Learning accepted to AAAI 2023. To be co-presented with Rajeev Alur, Osbert Bastani, and Kishor Jothimurugan.
- New paper Compositional Safety LTL Synthesis accepted at VSTTE 2022.
- FLoC time of the year. Excited to share plenty of new work!
- Presenting On the Unusual Effectiveness of Automata in Quantitative Reasoning at VardiFest 2022.
- Presenting Specification-Guided Learning of Nash Equilibria with High Social Welfare at CAV 2022.
- Presenting A Framework for Transforming Speicifcations in Reinforcement Learning at SYNT 2022.
- [Invited Speaker] Presenting Fomral Methods in Reinformcement Learning at WOLVERINE 2022.
- Roberto Cipollone will present our work on Compositional RL from Logical Specifications at PRL@IJCAI as none of the authors could make it to the workshop. Thank you, Roberto!
- Informally accepted Tenure-Track Assistant Professor position in the School of Computing at Georgia Insititute of Technology. The job market has concluded! :)