Pascal Sturmfels

Pascal Sturmfels

PhD Student in Computer Science

University of Washington

Biography

I am currently a PhD student at the University of Washington in the Baker Lab. I am currently interested in the intersection between deep learning and protein biology, specifically protein design. Much of my previous research focused on methods for explainable AI, specifically around feature attribution.

Interests
  • Machine Learning
  • Protein Design
  • Explainable AI
Education
  • PhD in Computer Science, Current

    University of Washington

  • B.E. in Computer Science, 2017

    University of Michigan

Recent Publications

Quickly discover relevant content by filtering publications.
(2021). Automated Brain Masking of Fetal Functional MRI with Open Data. Neuroinformatics.

PDF Cite Code

(2021). Explaining explanations: Axiomatic feature interactions for deep networks. Journal of Machine Learning Research.

PDF Cite Code

(2021). FoggySight: A Scheme for Facial Lookup Privacy.. Proc. Priv. Enhancing Technol..

PDF Cite Code

(2021). Improving performance of deep learning models with axiomatic attribution priors and expected gradients. Nature Machine Intelligence.

PDF Cite Code

Projects

Path Explanations

Path Explanations

A GitHub package to explain not only which features are most important to a neural network prediction, but also how those features interact with each other.

Attribution Priors

Attribution Priors

A GitHub package to regularize a neural network’s feature attributions during training in order to encode human-level prior knowledge into the training process.