Monte Carlo Dropout
PulseAugur coverage of Monte Carlo Dropout — every cluster mentioning Monte Carlo Dropout across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
New Physics-Informed Neural Network Tackles High-Dimensional Detection Problems
Researchers have developed QD-PINN, a novel framework utilizing Physics-Informed Neural Networks (PINNs) to address the challenges of Bayesian Quickest Detection problems in high-dimensional diffusion processes. This ap…
-
AI framework enhances structural health monitoring with synthetic data
Researchers have developed a novel data augmentation framework for structural health monitoring in aerospace applications. This framework utilizes conditional generative adversarial networks and Monte Carlo dropout to g…
-
New Contrastive Uncertainty Learning improves MRI segmentation accuracy
Researchers have developed a novel framework called Contrastive Uncertainty Learning (CUL) to improve the accuracy of anatomical structure segmentation in prostate MRIs. CUL utilizes predictive uncertainty to guide repr…
-
New framework improves data efficiency in engineering inverse design
Researchers have developed a new framework called CoNN-AL to address the high cost of data acquisition in engineering inverse design problems. This framework integrates stream-based active learning with a Cooperative Ne…
-
Connectome Graph Learning Needs Uncertainty Quantification for Trustworthy Biomarkers
A new paper published on arXiv explores the critical need for uncertainty quantification (UQ) in graph learning models applied to connectomics. The research highlights that while models like Graph Attention Networks (GA…
-
AI uncertainty quantification evaluated for trustworthy glioma diagnosis
Researchers have developed a task-aware evaluation strategy for uncertainty quantification (UQ) in AI systems designed for glioma diagnosis using MRI data. The study focused on a multi-task deep learning framework that …
-
Federated Learning Framework Enhances Infant Movement Analysis with Uncertainty Awareness
Researchers have developed a novel federated learning framework for analyzing infant movements, addressing privacy concerns in clinical settings. This framework utilizes Uncertainty-Aware Federated Averaging (UA-FedAvg)…
-
New foundation model IPM-FM targets industrial process monitoring
Researchers have developed IPM-FM, a novel foundation model designed for industrial process monitoring. This model leverages self-supervised pretraining on unlabeled industrial data to learn general representations, whi…
-
ToxLens framework enhances molecular toxicity prediction with leakage-aware learning
Researchers have developed ToxLens, a novel graph-learning framework designed to improve the accuracy and reliability of molecular toxicity predictions. This framework addresses the issue of performance overstatement in…
-
AI weather model Aardvark goes probabilistic, attributing forecast uncertainty
Researchers have developed a probabilistic version of the Aardvark Weather model, an end-to-end AI system for weather forecasting. By incorporating learned noise in the observation encoder and Monte Carlo Dropout in the…
-
New research explores uncertainty quantification and lightweight models for semantic segmentation
Researchers are exploring methods to improve the reliability and robustness of semantic segmentation models, particularly for safety-critical applications. One paper investigates the integration of uncertainty quantific…
-
AI model enhances diabetic retinopathy grading with uncertainty awareness
Researchers have developed a new pipeline for automated diabetic retinopathy (DR) grading that incorporates lesion-aware preprocessing, ordinal predictions, and uncertainty estimation. The system uses a specific feature…
-
New research questions Fréchet Inception Distance trustworthiness
A new paper published on arXiv explores the trustworthiness of the Fréchet Inception Distance (FID) metric, commonly used to evaluate synthetic image quality. The research, authored by Ciaran Bench, investigates how sto…
-
AI uncertainty metrics fail for lung nodule presence ambiguity
A new arXiv paper investigates the effectiveness of aleatoric uncertainty estimation in deep learning for 3D lung nodule segmentation. The study found that standard entropy-based uncertainty measures, while correlating …
-
New method ProteoKnight uses image encoding for phage protein classification
Researchers have developed ProteoKnight, a novel image-based encoding method for classifying Phage Virion Proteins (PVP). This technique adapts the DNA-Walk algorithm to capture intricate protein features, achieving 90.…
-
New Bayesian Explanation Method Enhances Power Quality Disturbance Classifier Reliability
This paper introduces a novel post-hoc Bayesian explanation method for deep learning classifiers used in power quality disturbance recognition. The method employs a Laplace approximation to efficiently derive an approxi…
-
New ensemble deep learning framework enhances skin lesion classification accuracy
Researchers have developed a new deep learning framework for classifying skin lesions from medical images. This framework combines a vision transformer model, MaxViT-Tiny, with two convolutional neural network models, C…
-
BERT-based QA models assessed for reliability; RoBERTa shows most stability
A new study published on arXiv evaluates the reliability of several BERT-based models, including RoBERTa, ALBERT, and DistilBERT, when applied to question-answering tasks. Researchers assessed model stability by introdu…
-
New AI framework enhances skin lesion classification with uncertainty and explainability
Researchers have developed a new framework for classifying skin lesions that combines deep ensemble learning with uncertainty quantification and explainability techniques. This approach uses multiple models, including v…
-
LLM uncertainty quantification research explores calibration for reliable answers
Two research papers explore methods for improving the reliability of answers generated by large language models (LLMs), particularly in question-answering tasks. The first paper introduces A-CRC-QA, a post-hoc calibrati…