Training Large-Scale Deep Nets with RL with Nando de Freitas - TWiML Talk #213

This article summarizes a podcast episode featuring Nando de Freitas, a DeepMind scientist, discussing his research on artificial general intelligence (AGI). The focus is on his team's work presented at NeurIPS, specifically papers on using YouTube videos to train agents for hard exploration games and one-shot high-fidelity imitation learning for training large-scale deep nets with Reinforcement Learning (RL). The article highlights the intersection of neuroscience and AI, and the pursuit of AGI through advanced RL techniques. The episode likely delves into the specifics of these papers and the challenges and advancements in the field.