Javier González

About

I am a Principal Researcher in Machine Learning in the Biological NLP/Real World Evidence group at Health Futures. I work in machine learning methods for healthcare. My research focuses on the following areas:

• Uncertainty quantification and trustworthy machine learning for healthcare: in the healhcare domain it is critical to develop methods that we can trust and that are safe to use. A basic behaviour for machine learning models to be trustworthy is to indentify scenarios where they are not equipped to provide and answer. Models should “know when they don’t know“. I work on probabilistic machine learning methods that can properly quantify uncertainy and that can be trusted for decision making in critical domains.
• Causal inference and Real-world evidence (RWE) for precision medicine: Randomised control trials (RCTs) are the gold standard for studying causal relationships between treatments and outcomes. They are, however, expensive, and time-consuming, while covering only a limited pool of patients. On the other hand, other sources of real-world data (RWD) routinely collected from medical practice, like electronic medical records, provide a rich and accesible source information that can augment RCT. A core area of my researh is to develop causal methods for RWD, which have the potential to provide patient-level insights and other real-world evidence useful to improve precision medicine.