Neural Networks and Brains

Explore a 48-minute conference talk from Devoxx that delves into the application of deep learning in predicting lung cancer patient survival. Learn about a framework utilizing convolutional networks to analyze multi-scale cancer data and clinical parameters, creating high-dimensional representations of tumor phenotypes. Discover how the integration of clinical data enhances model performance and examine the evaluation process across three independent cohorts. Gain insights into innovative techniques such as noise introduction for improving model robustness against delineation errors and the concept of priming, which boosts performance when training and testing on different cohorts.