- Francesco Pinto is a DPhil candidate at the University of Oxford in the Torr Vision Group (TVG). He is supervised by Philip H.S. Torr, Victor A. Prisacariu, Atılım Güneş Baydin and Puneet K. Dokania. Since the beginning of his DPhil he has collaborated with researchers from Facebook AI, FiveAI, Trilium Technologies and the European Space Agency.
- His research focus is on making Machine Learning and Deep Learning systems reliable when applied in the wild. Indeed, state-of-the-art models achieve outstanding performance on test sets that are i.i.d. (independently and identically distributed) with respect to the training set, but their performance is brittle when this assumption does not hold. Furthermore, these models can produce wrong predictions while being extremely confident. This is hindering their application in safety-critical applications.
- His up to date list of publications is at Google Scholar, and a summary of his previous experiences is given in his CV (last update: 9 Feb 2022).
- His interests are in Deep Learning, Computer Vision, Space Applications, Probabilistic Programming, Bayesian Statistics, Gaussian Processes.
- His specialistic expertise is in image classifiers calibration, distribution shift robustness, misclassification detection, out-of-distribution samples detection.
- He acquired his Bachelor and Master Degrees in Computer Engineering at the Politecnico di Milano (Italy), obtaining his degree defending a thesis on image inpainting with the supervision of A. Romanoni and M. Matteucci.
- He graduated with Honours in Piano Playing at the Conservatory G. Martucci (Italy).
