Neuromancer: Differentiable Programming Library for Data-driven Modeling and Control

Explore a comprehensive talk on NeuroMANCER, an open-source differentiable programming library for solving parametric constrained optimization problems, physics-informed system identification, and parametric model-based optimal control. Delve into the library's PyTorch-based architecture, which integrates machine learning with scientific computing to create end-to-end differentiable models and algorithms embedded with prior knowledge and physics. Learn about the library's focus on research, rapid prototyping, and streamlined deployment, as well as its emphasis on extensibility and interoperability with the PyTorch ecosystem. Discover numerous tutorial examples demonstrating the use of physics-informed neural networks for solution and parameter estimation of differential equations, learning to optimize methods with feasibility restoration layers, and differentiable control algorithms for learning constrained control policies for nonlinear systems. Recorded on September 26th, 2023, this 59-minute presentation by Ján Drgoňa offers valuable insights into the capabilities and applications of the NeuroMANCER library.