Public articles linked to the same research event.
arXiv This work applies amortized structured stochastic variational inference to Gaussian process latent variable models so that the variational posterior over the latent space depends conditionally on the values of the inducing points, moving beyond the mean-field approximation between inducing points and latent variables, and reports that this more flexible variational posterior improves several metrics relating to the reconstruction of points on the data manifold.
This work applies amortized structured stochastic variational inference to Gaussian process latent variable models so that the variational posterior over the latent space depends conditionally on the values of the inducing points, moving beyond the mean-field approximation between inducing points and latent variables, and reports that this more flexible variational posterior improves several metrics relating to the reconstruction of points on the data manifold.
This work applies amortized structured stochastic variational inference to Gaussian process latent variable models so that the variational posterior over the latent space depends conditionally on the values of the inducing points, moving beyond the mean-field approximation between inducing points and latent variables, and reports that this more flexible variational posterior improves several metrics relating to the reconstruction of points on the data manifold.
This work applies amortized structured stochastic variational inference to Gaussian process latent variable models so that the variational posterior over the latent space depends conditionally on the values of the inducing points, moving beyond the mean-field approximation between inducing points and latent variables, and reports that this more flexible variational posterior improves several metrics relating to the reconstruction of points on the data manifold.