Willa Potosnak is a 2nd year Ph.D. student at the Carnegie Mellon University (CMU) Robotics Institute (RI).
She is currently researching methods to improve modeling of complex dynamical system with Deep Learning. In particular, she is interested in the intersection of mechanistic modeling and Deep Learning that incorporates fundamental principles, such as differential equations, from domains in models. This approach can not only enhance model performance but can also facilitate learning underlying mechanics of the system, enabling tailored interventions to mitigate adverse outcomes.
She has worked on several projects to improve outcome prediction using forecasting and survival analysis frameworks while contributing open-source code.
Research
Physics-Informed Deep Learning \ Mechanistic Modeling
Efforts to supplement deep learning models with first principles from domain area have demonstrated successful application in many domains such as pharmacology, biology, physics, and fluid dynamics, among others. The combination of domain-specific first principles with deep learning results in models that are not only data-driven but also consistent with the underlying physical, biological, or chemical processes. Furthermore, incorporating prior knowledge, For example, first principles may be in the form of systems of differential equations that aim to capture underlying mechanisms of data generation. In this case, deep learning paradigms learn parameters of the system while in turn being constrained by the system design to reflect only realistic modeling solutions. .
Forecasting
Time series healthcare data can be challenging to forecast since there are many exogenous factors that contribute to changes in signal trajectories or patterns. Incorporating exogenous information in models can improve forecasts and reduce variance in model predictions. Exogenous information, such as medications, are often represented as sparse variables in data, yet may have latent time-dependent effects not encapsulated in numeric data. I am currently researching Machine Learning approaches to develop and parametrize mathematical models that encapsulate relevant time-dependent information for use in improving forecasts for data with sparse features.
Publications and Presentations
A complete list of publications can be found here.
Forecasting Response to Treatment with Global Deep Learning and Patient-Specific Pharmacokinetic Priors
W. Potosnak, C. Challu, K. G. Oliveras, A. Dubrawski
We propose a novel hybrid global-local architecture and a pharmacokinetic encoder that informs deep learning models of patient-specific treatment effects. In the Machine Learning for Health (ML4H) Findings Track Collection.
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
auton-survival is 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.
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.
Contact
For questions regarding research, please reach out to the following email.