[text]

[text]

Pulled in from source.

I am broadly interested in theory-driven deep learning applications. 

Currently I'm working on improving generative models, particularly diffusion models, under the manifold hypothesis.

I completed my PhD at the Faculty of Computer Science at the Technion focused on deep learning and statistical methods for high-resolution medical images and molecular measurements. I then worked as a postdoctoral research assistant where I found myself drawn towards generative models due their strong theoretical backing. This led me to develop a new class of generative models -- Generative Topological Networks -- which I will continue to explore alongside other generative approaches. 

For the most up-to-date information, best to visit: https://alonalj.github.io/ or find me on linkedin.

[e.g. educational background, biographical details]

[areas of research interest]

I am an ERC-funded postdoctoral researcher in the Department of Statistics at the University of Oxford, working with Prof. Pier Palamara.  I recently completed my DPhil in Genomic Medicine and Statistics, jointly supervised by Prof. Simon Myers and Prof. Pier Palamara.  My research focuses on developing Bayesian methods for statistical and population genetics, with applications to ancestry decomposition and genome-wide association studies. Prior to this, I earned my bachelor’s and master’s degrees in Electrical Engineering from IIT Bombay, India.

• Bayesian machine learning techniques for genome-wide association studies (GWAS).
• Genome-wide genealogies to analyze past human history, particularly "ghost" populations
• Uncertainty-aware and truth-worthy machine learning.

Loya, H., Kalantzis, G., Cooper, F. et al. A scalable variational inference approach for increased mixed-model association power. Nat Genet 57, 461–468 (2025). https://doi.org/10.1038/s41588-024-02044-7

Dupont, E., Loya, H., Alizadeh, M., Goliński, A., Teh, Y. W., & Doucet, A. (2022). Coin++: Neural compression across modalities. arXiv preprint 
https://arxiv.org/abs/2201.12904

Bezeljak, U., Loya, H., Kaczmarek, B., Saunders, T. E., & Loose, M. (2020). Stochastic activation and bistability in a Rab GTPase regulatory network. Proceedings of the National Academy of Sciences, 117(12), 6540-6549.

Loya, H., Poduval, P., Anand, D., Kumar, N., & Sethi, A. (2020). Uncertainty estimation in cancer survival prediction. arXiv preprint 
https://arxiv.org/abs/2003.08573