George Stamatescu
Student Experience
Division of Academic and Student Engagement
I am interested in the problem of making sequential decisions under uncertainty, in particular when you don't know the system that you're attempting to control. Although simple situations like the multi-armed bandit are well understood, extending this theory to the case where the system has a dynamically evolving state is difficult: optimal solutions become computationally intractable and even principled heuristics from the bandit setting are difficult to apply. I am currently studying this problem using tools from statistical physics and field theory, which might help bring new insight into the "nature" of sequential decision problem.
I am interested in the problem of making sequential decisions under uncertainty, in particular when you don't know the system that you're attempting to control. Although simple situations like the multi-armed bandit are well understood, extending this theory to the case where the system has a dynamically evolving state is difficult: optimal solutions become computationally intractable and even principled heuristics from the bandit setting are difficult to apply. I am currently studying this problem using tools from statistical physics and field theory, which might help bring new insight into the "nature" of sequential decision problem.
-
Appointments
Date Position Institution name 2020 - ongoing Post-doctoral researcher University of Adelaide -
Education
Date Institution name Country Title 2014 - 2019 University of Adelaide Australia PhD -
Research Interests
-
Journals
Year Citation 2018 Stamatescu, G., White, L., & Bruce-Doust, R. (2018). Track Extraction with Hidden Reciprocal Chains. IEEE Transactions on Automatic Control, 63(4), 1097-1104.
Scopus18 WoS17 -
Conference Papers
Year Citation 2020 Stamatescu, G., Gerace, F., Lucibello, C., Fuss, I., & White, L. (2020). Critical Initialisation in Continuous Approximations of Binary Neural Networks. In Proc ICLR 2020 (pp. 1-23). online: ICLR. 2015 Stamatescu, G., Dick, A., & White, L. (2015). Multi-camera tracking of intelligent targets with Hidden Reciprocal Chains. In Proceedings of the 2015 International Conference on Digital Image Computing: Techniques and Applications (pp. 1-8). Adelaide, Australia: IEEE.
Scopus5
Maths Learning Centre Tutor (2012- present).
Control III Tutor (2014).
-
Past Higher Degree by Research Supervision (University of Adelaide)
Date Role Research Topic Program Degree Type Student Load Student Name 2021 - 2024 Co-Supervisor Analysis of New Methods for Inference in Markov Decision Processes Master of Philosophy Master Full Time Mr Blake Edward Donnelly
Connect With Me
External Profiles