Software

Selected projects:

• Conformalized learning: uncertainty-aware training of deep neural network classifiers.
• CHR: powerful and adaptive conformal prediction intervals for regression.
• Conformal p-values: software for computing calibration-conditional conformal p-values for outlier detection.
• KnockoffGWAS: GWAS method localizing causal variants across the genome, accounting for population structure and familial relatedness (aka KnockoffZoom v2).
• ARC: Python package for adaptive and reliable classification.
• KnockoffZoom: Flexible tool for the multi-resolution localization of causal variants across the genome.
• DeepKnockoffs: Python package for approximate knockoffs and model-free variable selection.
• SNPknock: R package for generating knockoffs of hidden Markov models and genetic data.
• SNPknock for Python: Python package for generating knockoffs of hidden Markov models and genetic data. Note: this package is not up-to-date compared to the corresponding R package; the R package is currently recommended.
• Knockoffs: R and MATLAB packages for powerful and versatile controlled variable selection.

A full list of projects can be found on GitHub.