Learning High-Level Representations
Connor and Yannic discuss the importance of generative models in understanding and disentangling factors in data. They explore the concept of reconstructing reality through latent variables to achieve high-level representations, contrasting it with traditional classifiers. The conversation delves into the idea of aligning a latent model with visual input for object recognition.In this clip
From this podcast
Machine Learning Street Talk (MLST)
ICLR 2020: Yoshua Bengio and the Nature of Consciousness
Related Questions