kalman-jax
A JAX-based framework for approximate inference in Markov Gaussian processes using iterated Kalman smoothing.
What is kalman-jax?
Kalman-JAX is a Python library for approximate inference in Markov Gaussian processes using iterated Kalman smoothing algorithms. It provides efficient, linear-time inference methods for temporal Gaussian process models with non-conjugate likelihoods by leveraging state-space representations and JAX's computational capabilities.
Researchers and practitioners in probabilistic machine learning working with temporal data, Gaussian processes, or state-space models who need scalable inference methods with automatic differentiation support.
It uniquely combines classical state-space Kalman methods with modern JAX infrastructure, offering GPU/TPU acceleration, JIT compilation, and autodiff through inference loops for efficient and differentiable probabilistic modeling.
Overview
Approximate inference for Markov Gaussian processes using iterated Kalman smoothing, in JAX
Use Cases
Best For
- Temporal Gaussian process regression with non-Gaussian likelihoods
Related Projects
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
