Advances in Gaussian Process Modeling

The field of Gaussian process modeling is experiencing significant growth, with researchers exploring new techniques to improve scalability, efficiency, and accuracy. One notable trend is the development of approximate inference methods, which enable the application of Gaussian processes to large-scale datasets. Another area of focus is the integration of Gaussian processes with other machine learning frameworks, such as variational autoencoders, to leverage their strengths and improve overall performance. These advancements have far-reaching implications for various applications, including power systems, climate modeling, and Bayesian optimization. Noteworthy papers include:

  • Turbocharging Gaussian Process Inference with Approximate Sketch-and-Project, which presents an efficient algorithm for solving large-scale linear systems.
  • Bidirectional Information Flow, which introduces a hierarchical Gaussian process framework that establishes bidirectional information exchange between parent and child models.
  • Neighbour-Driven Gaussian Process Variational Autoencoders for Scalable Structured Latent Modelling, which proposes a neighbour-driven approximation strategy for scalable GPVAE inference.

Sources

Clustering Rooftop PV Systems via Probabilistic Embeddings

Bidirectional Information Flow (BIF) -- A Sample Efficient Hierarchical Gaussian Process for Bayesian Optimization

Turbocharging Gaussian Process Inference with Approximate Sketch-and-Project

Improving the Predictability of the Madden-Julian Oscillation at Subseasonal Scales with Gaussian Process Models

Gaussian Processes in Power Systems: Techniques, Applications, and Future Works

Bidirectional Variational Autoencoders

Neighbour-Driven Gaussian Process Variational Autoencoders for Scalable Structured Latent Modelling

Built with on top of