About

Willa Potosnak is a incoming Robotics Ph.D. student at the Carnegie Mellon University (CMU) Robotics Institute (RI). She is currently researching methods to improve machine learning (ML) model integrity with the objective of ensuring reliable implementation of artificial intelligence (AI) in practice. Among her research goals are to improve ML methods used to aid diagnostics and enhance people’s quality of life.

Research Intern in the Auton Lab

Willa currently works in the Auton Lab at Carnegie Mellon University under the mentorship of Dr. Artur Dubrawski and Dr. Kyle Miller.

Resume

Complete resume can be found here.

Education

Ph.D., Robotics

Carnegie Mellon University, Robotics Institute, Pittsburgh, PA

Beginning Fall 2022

B.S., Biomedical Engineering with Minor in Mathematics

Duquesne University, Rangos School of Health Sciences, Pittsburgh, PA

2018 - 2022
  • Summa Cum Laude

Research Experience

Machine Learning Research Intern

Auton Lab, Carnegie Mellon University, Pittsburgh, PA

August 2021 - Present
  • Conducting research with machine learning (ML) to improve the reliability of algorithms for distributed intra- and extra-institutional organizations with data privacy constraints.
August 2020 - May 2021
  • Advanced research with ML to develop intra-operative binary classification models to predict multiple post-operative outcomes for cardiothoracic surgery patients.

Robotics Institute Summer Scholar

Auton Lab, Carnegie Mellon University, Pittsburgh, PA

Summer 2021
  • Conducted research with ML to improve the reliability of algorithms that aim to provide insight into beneficial knowledge transfer opportunities.
Summer 2020
  • Conducted research with ML to develop an intra-operative binary classification model to predict acute renal failure for cardiothoracic surgery patients.

Projects

Projects include both research as well as volunteer work.

Post-Cardiac Surgery Outcome Prediction

The objective of this research effort is to improve individual patient risk predictions for multiple outcomes, such as mortality and renal failure, following cardiac-surgeries, including coronary artery bypass grafting (CABG). The broader inclusion of time-series intraoperative data in risk models to improve post-operative predictions has been one focus of the research effort.

RoboticsEd

RoboticsEd is a website of STEM resources developed with the goal of making robotics education more accessible to educators and their students. I am currently a volunteer co-lead of resource content and website design for a team of students from the robotics community. This initiative is co-lead by Rachel Burcin and Dr. Maxim Likhachev.

Publications and Presentations

A complete list of publications can be found here.

auton-survival: an Open-Source Package for Regression, Counterfactual Estimation, Evaluation and Phenotyping with Censored Time-to-Event Data

W. Potosnak, C. Nagpal, A. Dubrawski

We present auton-survival, an open-source repository of tools for survival regression, adjustment in the presence of domain shift, counterfactual estimation, phenotyping for risk stratification as well as estimation of treatment effects with censored time-to-event data. In Proceedings of the 7th Machine Learning for Healthcare Conference, PMLR.

Paper  Blog  Code 

Machine Learning Models with Intraoperative Features Improve Risk Predictions Following CABG

W. Potosnak, K. A. Dufendach, A. Kaczorowski, J. K. Miller, A. Dubrawski

Risk models that identify patients likely to develop adverse outcomes following coronary artery bypass grafting (CABG) rely mainly on preoperative data. Machine Learning was used to determine whether inclusion of intraoperative data offers improved model risk predictions for multiple outcomes following CABG. Presented at the 2022 Society of Thoracic Surgeons Coronary Conference.

Abstract  Conference Presentation Agenda

Robust Rule Learning for Reliable and Interpretable Insight into Expertise Transfer Opportunities

W. Potosnak, S. Caldas, K. A. Dufendach, J. K. Miller, A. Dubrawski

We propose an algorithmic rule selection approach which aims to select a short list of human-interpretable rules that reliably identify subpopulation beneficiaries of knowledge transfers in the form of machine learning risk models. In Proceedings of the Thirty-Sixth AAAI Conference on Artificial Intelligence.

3-minute Video

Robust Interpretable Rule Learning to Identify Expertise Transfer Opportunities in Healthcare.

W. Potosnak, S. Caldas, K. A. Dufendach, J. K. Miller, A. Dubrawski

We aim to identify opportunities for beneficial knowledge transfers across healthcare sites using a proposed algorithmic rule selection approach that yields reliable and human-interpretable decision lists. Presented at the 2021 NeurIPS workshop, Bridging the Gap: From Machine Learning Research to Clinical Practice.

Graph Neural Networks for Improved El Nino Forecasting.

S. Rühling Cachay, E. Erickson*, A. F. C., Bucker*, E. Pokropek*, W. Potosnak*, S. Osei, B. Lütjens.

We propose the application of spatiotemporal Graph Neural Networks (GNN) to forecast El Niño–Southern Oscillation (ENSO) at longer lead times with finer granularity and improved predictive skill than current state-of-the-art methods. Presented at the 2020 NeurIPS workshop, Tackling Climate Change with Machine Learning.

Paper  Recorded Talk

Contact

For questions regarding research, please reach out to the following email.