Efficient Graph-Based Active Learning with Probit Likelihood via Gaussian Approximations

We present a novel adaptation of active learning to graph-based\nsemi-supervised learning (SSL) under non-Gaussian Bayesian models. We present\nan approximation of non-Gaussian distributions to adapt previously\nGaussian-based acquisition functions to these more general cases. We develop an\nefficient rank-one update for applying "look-ahead" based methods as well as\nmodel retraining. We also introduce a novel "model change" acquisition function\nbased on these approximations that further expands the available collection of\nactive learning acquisition functions for such methods.\n

stat.ML cs.LG Machine Learning and Algorithms Gaussian Processes and Bayesian Inference Bayesian Modeling and Causal Inference

View graph Source

Paper

Similar papers

Computationally Efficient Active Learning of Gaussian Graphical ModelsAbrar Zahin, Gautam Dasarathy2024 In Library

Active Sampling over Graphs for Bayesian Reconstruction with Gaussian EnsemblesKonstantinos D. Polyzos, Qin Lu, Georgios B. Giannakis2022 In Library

Scalable Bayesian Structure Learning for Gaussian Graphical Models Using Marginal Pseudo-likelihoodReza Mohammadi, Marit Schoonhoven, Lucas Vogels et al.2023 · arXiv (Cornell University) In Library

Bayesian Semi-supervised Learning with Graph Gaussian ProcessesYin Cheng Ng, Ricardo Silva2018 · Neural Information Processing Systems · 93 citations In Library

Solving Bayesian Inverse Problems With Expensive Likelihoods Using Constrained Gaussian Processes and Active LearningMaximilian Dinkel, Carolin M. Geitner, Gil Robalo Rei et al.2023 · arXiv (Cornell University) · 11 citations In Library

Active learning by increasing model likelihood for Gaussian mixture models based classifiersBambang Heru Iswanto2021 · IOP Conference Series Materials Science and Engineering In Library

Ensemble Gaussian Processes for Online Learning Over Graphs With Adaptivity and ScalabilityKonstantinos D. Polyzos, Qin Lu, Georgios B. Giannakis2021 · IEEE Transactions on Signal Processing · 23 citations In Library

Query Efficient Posterior Estimation in Scientific Experiments via Bayesian Active LearningKirthevasan Kandasamy, J. Schneider, B. Póczos2017 · Artificial Intelligence · 31 citations In Library

Mixture of Gaussian Processes for Bayesian Active LearningChristoffer Riis, Filipe Rodrigues, Francisco Camara Pereira2023 In Library

Observation-dependent Bayesian active learning via input-warped Gaussian processesSanna Jarl, Maria Bånkestad, Jonathan J. S. Scragg et al.2026 · arXiv.org · 0 citations In Library

Source of record

arXiv

44B

Welcome.

This panel sticks with you. Pick anything below and it opens right beside it, so you can dig through 60,000-plus records without ever losing your spot here.

Travel 44B

LibraryPapers, policy, standards, statute LabsEvery organization building AI ModelsIntelligence, price, and speed BenchmarksThe evaluation catalog IncidentsAI failures and harms SearchOne field across everything DashboardThe state of AI accountability in NY 44B RegistryThe Art. 44-B compliance portal