Master or bachelor

• Prediction-powered inference: valid statistics with machine learning predictions [30|40|30]
• Stress-testing zero-shot time series foundation models [20|60|20]
• One pretrained model for every dataset? Understanding and stress-testing TabPFN [20|60|20]
• Warp-speed bootstrap in Monte Carlo experiments [40|60|0]
• The cheap bootstrap [40|50|10]
• Fast permutation tests for variable importance in machine learning models [20|60|20]
• Benchmarking state-of-the-art p-value aggregation methods (with Alberto Fernández-de-Marcos) [30|40|30]
• Testing for gene expression rhythmicity (with Yolanda Larriba) [30|40|30]

Master only

• Testing by betting: e-values and anytime-valid inference [40|40|20]
• Score matching for non-Euclidean data [50|40|10]
• Inference with discrete data via score matching [40|40|20]
• Statistical inference via Stein's method [60|30|10]

Bachelor only

• The harmonic mean p-value in multiple testing problems [40|40|20]
• Score matching vs. maximum likelihood [50|40|10]

Here is an incomplete list of some master and bachelor theses that I have supervised. The list gives an idea of potential topics I would be interested in exploring under new perspectives.

• A deep dive into Kolmogorov–Arnold networks [30|40|30]
• Neural networks for density learning [20|70|10]
• Learning circadian times of gene expressions (with Yolanda Larriba) [20|50|30]
• Nonparametric estimation of the spherical regression function with spatially correlated errors (with Andrea Meilán-Vila) [60|30|10]
• Generative modeling through diffusions [30|50|20]
• Density estimation via normalizing flows [30|60|10]
• Two interdisciplinary thesis proposals in aerospace engineering within the INFLUENTIA project (with Marco Raiola) [10|30|60]
• Inferring the optimal exercise boundary of an American option (with Abel Guada Azze) [20|50|30]
• Optimal stopping of an Ornstein–Uhlenbeck bridge, with applications in mathematical finance (with Abel Guada Azze) [20|50|30]