Keenan Crane Compares Procedural Generators to Machine Learning

Keenan Crane, an Associate Professor of Computer Science and Robotics at CMU, posted a reply drawing a connection between classical procedural generators and contemporary machine learning. He observes that the spirit of these generators aligns with ML by drawing samples from a distribution based on user-defined parameters. Additionally, Crane points out that classic distributions can function in an autoregressive manner, depending on previous generation steps. The statement comes directly from the researcher's public reply.