Explore discriminative learning approaches for generative models in this 70-minute lecture delivered by Tony Jebara from Columbia University at Johns Hopkins University's Center for Language & Speech Processing. Learn how to bridge the gap between generative and discriminative machine learning paradigms by applying discriminative training criteria to generative model architectures. Discover techniques for optimizing generative models using discriminative objectives, understand the theoretical foundations behind this hybrid approach, and examine practical applications where discriminative learning can improve the performance of traditional generative models in classification and pattern recognition tasks.