Aleksandar Botev
Papers
1
Total Citations
4
H-Index
1
About
Aleksandar Botev is a leading researcher in probabilistic machine learning, with a primary focus on learning latent dynamics, Bayesian inference, and scalable deep learning. His work addresses the critical challenge of enabling ML systems to reason about complex physical systems from high-dimensional observations, such as images, without direct access to underlying state variables. Botev’s most cited paper, "Which priors matter? Benchmarking models for learning latent dynamics" (2021), provides a rigorous framework for evaluating how different prior assumptions influence the performance of latent dynamics models—a key consideration for applications in robotics and autonomous driving. This benchmarking study has become a foundational reference for researchers seeking to design more robust and interpretable models. Beyond this, Botev has contributed to advancing variational inference and neural network calibration, helping bridge the gap between theoretical probabilistic modeling and practical deployment. His work consistently emphasizes the importance of uncertainty quantification, making his research highly influential for students and practitioners working on safe and reliable AI systems.
Research Focus
Key Achievements
Top Papers
- 1Which priors matter? Benchmarking models for learning latent dynamics4 citations · 2021