Amortized Structured SVI for GP-LVMs
What it is
This paper proposes a method to improve Gaussian Process Latent Variable Models by using Amortized Structured Stochastic Variational Inference. It allows the variational posterior for latent variables to be conditionally dependent on inducing points, overcoming the limitations of standard mean-field approximations.
The problem it solves
Standard mean-field variational approximations in GP-LVMs limit the effectiveness of uncertainty estimation by assuming independence between inducing points and latent variables.
What you could use it for
- Improving uncertainty estimation in manifold learning
- Enhancing reconstruction accuracy for data lying on lower-dimensional manifolds
- Refining variational inference in Gaussian Process models
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.