Bayesian Neural Networks
PulseAugur coverage of Bayesian Neural Networks — every cluster mentioning Bayesian Neural Networks across labs, papers, and developer communities, ranked by signal.
5 day(s) with sentiment data
-
Photonic AI uses light speckle for uncertainty-aware Bayesian networks
Researchers have developed a novel photonic processor that utilizes temporal speckle to represent probabilistic weights for uncertainty-aware artificial intelligence. This approach encodes mean and variance directly int…
-
AI drones shape Ukraine war; BNNs analyzed in new paper
A new paper on arXiv explores the theoretical underpinnings of Bayesian Neural Networks (BNNs) in statistical decision theory, focusing on their optimality under quadratic and Kullback--Leibler loss. Separately, the use…
-
New method enables sparse Bayesian neural networks with exact zero-weight sampling
Researchers have developed a novel method for creating sparse Bayesian neural networks (BNNs) using sticky piecewise deterministic Markov processes (PDMPs). This approach directly incorporates pruning within a Bayesian …
-
New Bayesian Neural Networks Enhance Deep Reinforcement Learning Exploration
Researchers have developed a novel class of Bayesian neural networks (BNNs) that place Dirichlet Process (DP) priors directly on the data-generating distribution, rather than on network parameters. This approach aims to…
-
New method accelerates Bayesian Neural Networks with single forward pass
Researchers have developed a new method to accelerate Bayesian Neural Networks (BNNs) by using a single Probabilistic Forward Pass (PFP) and code generation. This approach significantly reduces the computational cost as…
-
Bayesian Neural Networks Achieve Adaptive Inference with Statistical Guarantees
Researchers have developed a novel method using confidence sequences to dynamically determine the necessary number of Monte Carlo samples for Bayesian neural network predictions. This approach ensures statistical guaran…
-
Bayesian neural networks: Likelihood tempering impact analyzed
Researchers have investigated the impact of likelihood tempering on variational Bayesian linear neural networks. They found that in wide networks, Gaussian mean-field variational inference can lead to "prior dominance,"…
-
New GMVIP method enhances Bayesian inference for implicit-process priors
Researchers have introduced Generalized Matheron Variational Implicit Processes (GMVIP), a novel variational family designed for posterior inference with implicit-process priors. These priors, often defined through mech…
-
New framework certifies safety in federated Bayesian learning models
Researchers have developed a new framework for certifying the safety of Bayesian neural networks in federated learning scenarios. This method, called Posterior Event Transport, addresses the challenge that local safety …
-
New ensemble sampling methods promise improved efficiency and accuracy in ML research
Two new research papers propose novel ensemble sampling techniques to improve the efficiency and accuracy of model exploration in machine learning. The first paper, "Linear Ensemble Sampling with Smaller Ensembles," int…
-
New HyperMC framework optimizes SGMCMC hyperparameters
Researchers have developed HyperMC, a novel framework for optimizing hyperparameters in Stochastic Gradient Markov Chain Monte Carlo (SGMCMC) methods. This approach utilizes a multi-fidelity tuning strategy, combining H…
-
New concept LPA explains uncertainty reduction in GNNs
Researchers have introduced Latent-Posterior Alignment (LPA), a new concept explaining how predictive uncertainty is reduced in Graph Neural Networks (GNNs) with Bayesian output layers. This phenomenon occurs as latent …
-
New AI methods optimize 3D printing quality under uncertainty · 3 sources tracked
Researchers have developed new methodologies for optimizing the fused filament fabrication (FFF) process, focusing on improving part quality under uncertainty. One approach uses Bayesian neural networks to predict geome…
-
New research explores uncertainty quantification in deep learning for diverse applications
Three new research papers explore advanced techniques for uncertainty quantification in deep learning models. The first paper introduces intuitionistic fuzzy deep randomized neural networks (IF-dRVFL and IF-edRVFL) to i…
-
AI advances photonic component design with neurosymbolic and BNN approaches
Researchers have developed new methods for designing photonic components using AI. One approach, "Constrained Co-Design for Photonic Bayesian Neural Networks," focuses on improving the uncertainty estimation of Bayesian…
-
New survey details uncertainty quantification for trustworthy deep learning
A new survey paper published on arXiv details methods for uncertainty quantification in deep learning, focusing on techniques relevant for trustworthy AI in safety-critical applications. The paper categorizes approaches…
-
Student's t-distribution outperforms Gaussian in Bayesian Neural Networks
Researchers have explored the impact of different likelihood distributions on the performance of Bayesian Neural Networks (BNNs). While Gaussian distributions are commonly used for modeling uncertainty in BNNs due to co…
-
Neural network generalization near interpolation analyzed via statistical mechanics · 2 sources tracked
Two new arXiv papers explore the behavior of shallow neural networks with extensive width, focusing on their generalization capabilities near the interpolation threshold. The research analyzes these networks using stati…
-
New method disentangles model and human uncertainty in facial age estimation
Researchers have developed a method to distinguish between model uncertainty and human data uncertainty in facial age estimation tasks. By training Bayesian Neural Networks on the APPA-REAL dataset with varying data siz…
-
New SCROLL method optimizes Bayesian neural networks via Bethe free energy
Researchers have developed a new method for training Bayesian neural networks called SCROLL (Shared-Cavity fRee-rOuting Last-Layer). This approach optimizes the Bethe free energy rather than the typical evidence lower b…