Monte Carlo Dropout

Monte Carlo Dropout is a technique used to estimate uncertainty in the predictions of neural networks, essentially treating dropout during training as a Bayesian approximation. Current research focuses on improving the accuracy and interpretability of uncertainty estimates, particularly within specific model architectures like U-Nets and through the use of symmetric regressors, and on developing efficient implementations for resource-constrained environments such as FPGAs and spiking neural networks. This approach is significant because reliable uncertainty quantification is crucial for deploying neural networks in high-stakes applications like medical diagnosis and autonomous driving, where understanding the confidence of predictions is paramount.

Papers

June 21, 2024

Bayesian neural networks for predicting uncertainty in full-field material response
George D. Pasparakis, Lori Graham-Brady, Michael D. Shields
High Uncertainty Anticipation Bayesian Neural Network 3D Microstructures Solid Mechanic Stress Field Monte Carlo Dropout

April 18, 2024

A Symmetric Regressor for MRI-Based Assessment of Striatal Dopamine Transporter Uptake in Parkinson's Disease
Walid Abdullah Al, Il Dong Yun, Yun Jung Bae
Common Disease Parkinson Disease Multi Parametric MRI Monte Carlo Dropout

August 24, 2023

Single-shot Bayesian approximation for neural networks
Kai Brach, Beate Sick, Oliver Dürr
Neural Network Deep Neural Network Deep Ensemble Dropout Layer Monte Carlo Dropout Moment Propagation

August 13, 2023

When Monte-Carlo Dropout Meets Multi-Exit: Optimizing Bayesian Neural Networks on FPGA
Hongxiang Fan, Hao Chen, Liam Castelli, Zhiqiang Que, He Li, Kenneth Long, Wayne Luk
Bayesian Neural Network Field Programmable Gate Array Multi Exit Bayes Net Monte Carlo Dropout

April 20, 2023

Efficient Uncertainty Estimation in Spiking Neural Networks via MC-dropout
Tao Sun, Bojian Yin, Sander Bohte
Spiking Neural Network Uncertainty Estimation Neuromorphic Hardware Predictive Uncertainty Structured Dropout Classical Neural Network Efficient Uncertainty Estimation Monte Carlo Dropout

October 27, 2022

Adapting Neural Models with Sequential Monte Carlo Dropout
Pamela Carreno-Medrano, Dana Kulić, Michael Burke
Neural Model Particle Filter Monte Carlo Dropout Dropout Mask

September 22, 2022

Beyond Voxel Prediction Uncertainty: Identifying brain lesions you can trust
Benjamin Lambert, Florence Forbes, Senan Doyle, Alan Tucholka, Michel Dojat
3D Medical Image Voxel Wise Brain Lesion Voxel Scale Uncertainty Monte Carlo Dropout

February 24, 2022

Embedded Ensembles: Infinite Width Limit and Operating Regimes
Maksim Velikanov, Roman Kail, Ivan Anokhin, Roman Vashurin, Maxim Panov, Alexey Zaytsev, Dmitry Yarotsky
Ensemble Model Model Ensembling Infinite Width Monte Carlo Dropout Dynamical Regime Independent Model Distinct Functional Regime