Explore autoregressive long-context music generation using Perceiver AR in this one-hour presentation by Cătălina Cangea from DeepMind. Delivered as part of the DataLearning working group's weekly meeting on October 4, 2022, the talk delves into innovative techniques for creating extended musical compositions through advanced machine learning models. Gain insights into the intersection of data assimilation and artificial intelligence as applied to music generation. Suitable for researchers, students, and professionals interested in the latest developments in AI-driven creative technologies and long-form content generation.