A pytorch implementation of MCDO(Monte-Carlo Dropout methods)

All the material needed to use MC-CP and the Adaptive MC Dropout method

Uncertainty Estimation Using Deep Neural Network and Gradient Boosting Methods

Bayesian deep learning experiments

An experimental Python package for learning Bayesian Neural Network.

PyTorch-native Bayesian neural networks for uncertainty quantification in deep learning

An NLP Model used for automated assignment of bug reports to the relevant engineering team. Utilizes a novel confidence bounding approach - Monte Carlo Dropout, and assigns underconfident predictions to a queue for human review. Built for Pegasystems Inc.

PyTorch implementation of landmark-based facial expression recognition using Spatio-Temporal BiLinear Networks (ST-BLN)

Hybrid CNN-Transformer model for automated exoplanet transit detection on NASA Kepler light curves. Features dual scale CNN branches, Transformer based sequence modelling, Grad CAM + SHAP + attention explainability, and Monte Carlo Dropout uncertainty quantification. 5 fold cross validated with baseline comparisons.

Hybrid Uncertainty Quantification for Bioactivity Assessment

Comparison of a network implemented via Variational Inference with the same network implemented via Monte Carlo Dropout

(Forked Version) Experiments used in "Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning"

GUI for analyzing deep learning models with uncertainty in predictions

Probabilistic approach to neural nets - modern scalable approximate inference methods

RAG system with Bayesian uncertainty quantification using Beta priors and Monte Carlo Dropout

Uncertainty-aware financial forecasting for the Borsa Istanbul 100 Index. Transformer model with Monte Carlo Dropout for epistemic uncertainty, integrating Turkish equity, US equity, and FRED-MD macro data. Coursework for COM0415 Applied Deep Learning.

monitor.ai: Non-intrusive monitoring for FDA-approved medical AI, helping healthcare organizations ensure safety and compliance without modifying validated models.

Waveform based ventricular tachycardia alarm validation using a 1D CNN + LSTM architecture with Monte Carlo Dropout for uncertainty estimation.

Loto6 lottery number prediction with BiLSTM + Monte Carlo Dropout (educational use)