Machine Learning for Personalized Medicine

Marie-Curie Action: "Initial Training Networks"

Project: Learning Decision Support for Personalized Medicine

Cristóbal Esteban
Volker Tresp
Munich, Germany

Cristóbal Esteban
was born in Spain. He studied Telecommunication Engineering and then he obtained an MSc in Logic and Artificial Intelligence working on the topic of Recommender Systems for his thesis.

Project description

The main goal of the project is to develop a decision support system for physicians based on large-scale datasets.
The project is developed in the context of precision medicine, since we use data to analyze the causes of an individual patient's disease and then to utilize targeted treatments to address that individual patient's disease process. The patient's response is then tracked as closely as possible and the treatment is finely adapted to the patient's response.

We are currently working with a database provided by the Charité University Hospital, Berlin, Germany, composed by patients that went through a kidney transplantation. This database contains approximately 5.000 patients whose evolution has been recorded during the last 30 years.
Our research is focused on issues as

  • assisting the physicians to make diagnosis and prescribe the most suitable medications for each case, and
  • helping them to decide which receiver is the best match for each donor.

Some of the challenges in which we are working to achieve our goals are:

  • Exploit learning decision support approaches that are suitable for noisy, incomplete and high dimensional medical and biomedical data with sparse relationships.
  • Exploit structured and unstructured patient data and background data from resources such as linked-life data including medical and biomedical ontologies.

Motivation for participating in the network

I am participating in the network to be able to work with the top experts in the field of Machine Learning and healthcare while I enrich the experience by sharing knowledge and projects with the fellows in the other nodes of the network.


Duration of fellowship: from July 2013 to July 2016


MLPM Publications:

- Denis Krompaß, Cristóbal Esteban, Volker Tresp, Martin Sedlmayr and Thomas Ganslandt. Exploiting Latent Embeddings of Nominal Clinical Data for Predicting Hospital Readmission. KI - Künstliche Intelligenz, December 2014.

- Cristóbal Esteban, Danilo Schmidt, Denis Krompaß, and Volker Tresp.Predicting Sequences of Clinical Events by using a Personalized Temporal Latent Embedding Model. Proceedings of the  IEEE International Conference on Healthcare Informatics  (ICHI), 2015.


- Cristóbal Esteban, Stephan Baier, Yinchong Yang, Danilo Schmidt, Denis Krompaß, and Volker Tresp. Representation Learning for Electronic Health Record Modeling. NIPS workshop on Machine Learning in Healthcare, 2015 (accepted).

- Volker Tresp, Cristóbal Esteban, Yinchong Yang, Stephan Baier, and Denis Krompass. Learning with Memory Embeddings. NIPS workshop on Nonparametric Methods for Large Scale Representation Learning, 2015 (accepted).

Like us on Facebook!