Deep Learning in Biomedicine: Challenges and Opportunities by Dr. Olivier Gevaert

Massive amounts of data are being generated in biomedicine ranging from molecular, to cellular, to tissue-scale data. It is now possible to study a patient from multiple angles and scales with multi-modal, multi-scale, high dimensional, and high throughput biomedical data. However, these data are noisy, sometimes parts are missing, and it is unknown what entities are important for “precision medicine,” the concept that medical care has to be designed to optimize treatment of a patient by using the patient’s own data. In parallel, deep learning has revolutionized fields such as image recognition, natural language processing and, more broadly, machine learning and AI. Now, deep learning is becoming increasingly more popular for analyzing biomedical data. This talk will introduce the types of data in biomedicine, including images, text to numeric data, and the potential opportunities and challenges for modeling this data using deep learning. Dr. Olivier Gevaert is an assistant professor of medicine (biomedical informatics) and biomedical data science at the Stanford University School of Medicine. The Gevaert Lab focuses on multi-scale biomedical data fusion focusing on applications in oncology. This involves modeling data at molecular, cellular, and tissue levels using a wide range of different technologies such as sequencing (DNA, RNA) and imaging (MRI, CT, PET, etc.). The lab develops computational approaches to find relevant patterns in these high dimensional, multi-modal data sets. These machine-learning methods combine elements from statistic and mathematics and apply them on multi-scale data, ranging from molecular biology, to pathology, to radiology.