Theofanis Karaletsos
Papers
1
Total Citations
7
H-Index
1
About
Theofanis Karaletsos is a leading researcher at the intersection of machine learning, probabilistic modeling, and reinforcement learning (RL). His work is distinguished by pioneering methods that enable AI systems to learn and adapt efficiently in complex, real-world environments. Karaletsos is best known for advancing transfer learning and online adaptation in continuous control, where he developed latent variable models that allow RL agents to generalize across subtle variations in physical dynamics—such as changes in mass or friction—without requiring retraining from scratch. This approach bridges the gap between sample efficiency and robustness, a critical challenge in deploying RL in robotics and autonomous systems. His highly cited 2018 paper on this topic has garnered significant attention for its practical impact. Beyond RL, Karaletsos has made notable contributions to deep generative models and Bayesian deep learning, often focusing on scalable inference and representation learning. His work consistently emphasizes principled, uncertainty-aware frameworks, making him a key figure in advancing robust, adaptable AI systems for real-world applications.
Research Focus
Key Achievements
Top Papers
- 1