This lecture discusses the challenges in humanoid reinforcement learning, such as learning whole-body motor skills and the difficulties in convergence due to biased samples. It also explores the application of assistive forces in humanoid robots to enhance their locomotion capabilities.