Georgios Tzannetos

I am a Doctoral student at the Max Planck Institute for Software Systems under the supervision of Dr. Adish Singla. I am interested in making Reinforcement Learning agents train more efficiently with fewer interactions with their environment. We get inspired by human learning and we design curriculum strategies that help agents learn faster. I hold a M.Sc degree from the Technical University of Munich and an M.Eng degree from the National Technical University of Athens. For more updated news you can check my personal website or follow me on Twitter.

Publications


2024

Reward Model Learning vs. Direct Policy Optimization: A Comparative Analysis of Learning from Human Preferences
International Conference on Machine Learning, ICML 2024.
A Nika, D Mandal, P Kamalaruban, G Tzannetos, G Radanović, A Singla
[arxiv] [pdf]
Proximal Curriculum with Task Correlations for Deep Reinforcement Learning
International Joint Conference on Artificial Intelligence, IJCAI 2024.
G Tzannetos, P Kamalaruban, A Singla
Learning Embeddings for Sequential Tasks Using Population of Agents
International Joint Conference on Artificial Intelligence, IJCAI 2024.
M Mahajan, G Tzannetos, G Radanovic, A Singla
[arxiv] [pdf] [code]
Neural Task Synthesis for Visual Programming
Transactions on Machine Learning Research, January 2024.
V Padurean, G Tzannetos, A Singla
[arxiv] [pdf] [code]

2023

Proximal Curriculum for Reinforcement Learning Agents
Transactions on Machine Learning Research, April 2023.
G Tzannetos, BG Ribeiro, P Kamalaruban, A Singla
[arxiv] [pdf] [code]