Myriad: a real-world testbed to bridge trajectory optimization and deep learning

TL;DR

Myriad is a JAX-based testbed that integrates trajectory optimization with deep learning for real-world continuous control problems, facilitating research in learning and planning in complex environments.

Contribution

It introduces a versatile testbed with real-world control environments and demonstrates a novel end-to-end learning approach using neural ODEs and implicit planning.

Findings

Enables machine learning practitioners to incorporate trajectory optimization.

Provides diverse real-world control environments for benchmarking.

Showcases a novel end-to-end learning and planning method with neural ODEs.

Abstract

We present Myriad, a testbed written in JAX for learning and planning in real-world continuous environments. The primary contributions of Myriad are threefold. First, Myriad provides machine learning practitioners access to trajectory optimization techniques for application within a typical automatic differentiation workflow. Second, Myriad presents many real-world optimal control problems, ranging from biology to medicine to engineering, for use by the machine learning community. Formulated in continuous space and time, these environments retain some of the complexity of real-world systems often abstracted away by standard benchmarks. As such, Myriad strives to serve as a stepping stone towards application of modern machine learning techniques for impactful real-world tasks. Finally, we use the Myriad repository to showcase a novel approach for learning and control tasks. Trained in a fully end-to-end fashion, our model leverages an implicit planning module over neural ordinary differential equations, enabling simultaneous learning and planning with complex environment dynamics.