Scalable Gaussian Process Structured Prediction for Grid Factor Graph Applications
نویسندگان
چکیده
Training: intractable likelihood likelihood needed for MCMC; grid MRF intractable due to global normalization. Solution: pseudo-likelihood approximation, locally normalized: p(y|f) = ∏P p=1 p(yp|yp+1 . . . yP , f) ≈ ∏P p=1 p(yp|neighbours(p), f) Prediction: intractable marginals Pixel-wise MAP intractable in grid MRF. Solution: Obtain approximate marginals using treereweighted belief propagation. Training: kernel matrix size How many factors f(xp,yp)? One for each p in Dtrain, yp ∈ Y, ie (10) × 10 = 10. So K square of shape 10 × 10. Solution: ensemble learning, split up training pixels over weak learners.
منابع مشابه
Gaussian process prediction for time series of structured data
Time series prediction constitutes a classic topic in machine learning with wide-ranging applications, but mostly restricted to the domain of vectorial sequence entries. In recent years, time series of structured data (such as sequences, trees or graph structures) have become more and more important, for example in social network analysis or intelligent tutoring systems. In this contribution, w...
متن کاملKernel Interpolation for Scalable Structured Gaussian Processes (KISS-GP)
We introduce a new structured kernel interpolation (SKI) framework, which generalises and unifies inducing point methods for scalable Gaussian processes (GPs). SKI methods produce kernel approximations for fast computations through kernel interpolation. The SKI framework clarifies how the quality of an inducing point approach depends on the number of inducing (aka interpolation) points, interpo...
متن کاملCollaborative Filtering with Side Information: a Gaussian Process Perspective
We introduce the Tucker Gaussian Process (TGP), a model for regression that regularises a Gaussian Process (GP) towards simpler regression functions for enhanced generalisation performance. We derive it using a novel approach to scalable GP learning, and show that our model is particularly well-suited to grid-structured data and problems where the dependence on covariates is close to being sepa...
متن کاملIt is all in the noise: Efficient multi-task Gaussian process inference with structured residuals
Multi-task prediction methods are widely used to couple regressors or classification models by sharing information across related tasks. We propose a multi-task Gaussian process approach for modeling both the relatedness between regressors and the task correlations in the residuals, in order to more accurately identify true sharing between regressors. The resulting Gaussian model has a covarian...
متن کاملStructured regression on partially observed evolving graphs with uncertainty propagation
Conditional probabilistic graphical models provide a powerful framework for structured regression in spatio-temporal datasets with complex correlation patterns. However, in real-life applications a large fraction of observations is often missing, which can severely limit the representational power of these models. We have proposed a Marginalized Gaussian Conditional Random Fields (m-GCRF) struc...
متن کامل