Watch a conference talk from TQC 2023 exploring quantum algorithms for training reinforcement learning policies through quantum interactions with an environment. Learn about the potential for quadratic speed-ups in sample complexity compared to classical approaches, particularly when using parameterized quantum circuits as policies. Discover how this research advances understanding of quantum computing's role in artificial intelligence by examining the power and limitations of quantum access to data in machine learning tasks. Follow along as the speaker covers key concepts including reinforcement learning scenarios, agent-environment interaction, policy gradient training, Monte Carlo estimation, quantum complexity and sampling. The presentation concludes with numerical results demonstrating the benefits of a fully-quantum reinforcement learning framework and includes an audience Q&A session.