Machine Learning-Guided Treatment Discovery

This lecture by the instructor covers the topic of personalized medicine and the Omics revolution, focusing on how machine learning can guide treatment discovery and planning. It discusses the use of high-throughput data to predict treatment outcomes, the limitations of traditional machine learning algorithms, and the development of robust models that encode problem nature as an inductive bias. The lecture also delves into predicting treatment responses using neural optimal transport and evaluating the CELLOT model. Future work includes spatiotemporal modeling of tissue structure and interaction, aligning models with high-throughput technologies, and quantifying therapy success.