Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only! Grab it Watch a 51-minute lecture from Johns Hopkins University professor Joshua Vogelstein at IPAM's Modeling Multi-Scale Collective Intelligences Workshop exploring how to adapt machine learning for dynamic, real-world environments. Discover a new theoretical framework called "prospective learning" that extends beyond traditional PAC learning to address evolving data distributions and goals over time. Learn why standard empirical risk minimization fails at certain basic prospective learning challenges and examine a novel prospective augmentation approach that offers solutions. Follow along with demonstrations showing how prospective learners outperform traditional methods on synthetic data and visual recognition tasks using MNIST and CIFAR datasets. Gain insights into improving AI systems for currently intractable problems while better understanding naturally intelligent systems that already solve these challenges.