Ph.D. Position in A.I. for healthcare

Your responsibilities

Our goal is to develop state-of-the-art approaches and build best-in-class methods to capitalize on multimodal clinical information for building robust decision support systems powered by (explainable and interpretable) predictive algorithms for guiding patient therapy across all disease stages, the assessment of treatment effects using counterfactual inference and the identification of causal mechanisms driving disease progression (see examples of our latest work 1-8 ).

Moreover, there are possibilities to expand the research scope towards working with heterogeneous biomedical data.

Your profile

Minimum qualifications :

• Master's degree (MSc) in computer science (with emphasis on machine learning), optimization, statistics, applied math or closely related discipline
• Proficient in Python and the scientific computing stack (SciPy, Numpy, Scikit- learn, pandas)
• Proficient in one of the deep learning frameworks (PyTorch, Tensorflow, or Jax)

Additional (preferred) qualifications :

• Knowledge of probabilistic graphical models (such as Gaussian processes, Neural Process, Bayesian inference or Bayesian NN)
• Knowledge of explainable AI methods (i.e., model explainability and interpretability)
• Knowledge / understanding of Generative ML models (i.e. Diffusion model, VAE, autoregressive language model, GAN, etc.)
• Experience using Linux systems and HPC infrastructure

What we offer

We offer an interdisciplinary research environment, the possibility to direct your own research and access to state-of-the-art computational resources infrastructure.

• Access to clinical datasets and medical expertise domain-knowledge (excellent medical doctors and research scientists)
• Ability to make a real and tangible impact in healthcare research
• Solve real-world problems and improve hospital-related processes and workflow
• Stimulating research environment and a place to grow academically and professionally
• Outstanding working conditions at the University of Zurich.

References

1 C. Trottet, A. Allam, A. N. Horvath, R. Micheroli, M. Krauthammer, and C. Ospelt, "Explainable Deep Learning for Disease Activity Prediction in Chronic Inflammatory Joint Diseases," in ICML 2023, IMLH Workshop, 2023.

Online . Available : https : / / openreview.net / forum?id W1y2ckWGuX

2 C. Trottet et al., "Explainable deep learning for disease activity prediction in chronic inflammatory joint diseases," medRxiv, 2023.

Online . Available : https : / / doi.org / 10.1101 / 2023.12.05.23299508 (accepted for publication in PLOS Digital Health)

3 C. Trottet et al., "Modeling Complex Disease Trajectories using Deep Generative Models with Semi-Supervised Latent Processes," in Machine Learning for Health (ML4H) 2023, 2023.

Online . Available : https : / / arxiv.org / abs / 2311.08149. DOI : https : / / doi.org / 10.48550 / arXiv.2311.08149

4 C. Trottet, M. Schürch, A. Mollaysa, A. Allam, and M. Krauthammer, "Generative time series models with interpretable latent processes for complex disease trajectories," in Deep Generative Models for Health Workshop NeurIPS 2023, 2023.

Online . Available : https : / / openreview.net / forum?id tiqs7trqcC

5 M. Schürch et al., "Generating Personalized Insulin Treatments Strategies with Deep Conditional Generative Time Series Models," in Machine Learning for Health (ML4H) 2023, 2023.

Online . Available : https : / / arxiv.org / abs / 2309.16521. DOI : https : / / doi.org / 10.48550 / arXiv.2309.16521

6 M. Schürch et al., "Generating Personalized Insulin Treatments Strategies with Conditional Generative Time Series Models," in Deep Generative Models for Health Workshop NeurIPS 2023, 2023.

Online . Available : https : / / openreview.net / forum?id SXw8DBKoRg

7 A. Allam et al., "Predicting Interstitial Lung Disease Progression in Patients with Systemic Sclerosis Using Attentive Neural Processes - A EUSTAR Study," medRxiv, 2024.

Online . Available : https : / / doi.org / 10.1101 / 2024.04.25.24306365

8 X. Zheng, M. Schürch, X. Chen, M. A. Komninou, R. Schüpbach, A. Allam, J. Bartussek, and M. Krauthammer, "Clustering of Disease Trajectories with Explainable Machine Learning : A Case Study on Postoperative Delirium Phenotypes," arXiv, eprint 2405.

03327, May 2024. Online . Available : https : / / arxiv.org / abs / 2405.03327. DOI : https : / / doi.org / 10.48550 / arXiv.2405.03327