ashish1610dhiman/UQ_deep_learning

UQ methods for DL

Introduction to Uncertainty Quantification for Deep Learning

Gaussian Process Regression for Yield curve

The codes/files accompanying the GP modelling of yield curve are given below:

• Download Yield Curve Data
• Yield curve data
• Modelling Yield curve with GP

Bayesian Regression Examples

This repository contains Jupyter Notebook files related to Bayesian regression problems on real estate dataset. These notebooks explore different techniques and approaches for Bayesian regression modeling, as given below:

• Part A (2 'x' variables): Bayesian Regression with SVI (Diagonal Covariance)
• Part B (6 'x' variables): Bayesian Regression with SVI (Diagonal Covariance)
• Part A: Bayesian Regression with SVI (Multi-dimensional HMC)
• Part B: Bayesian Regression with SVI (Multi-dimensional HMC)

Feel free to explore these notebooks to dive deeper into Bayesian regression and its various implementation techniques.

Quick Primer on UQ techniques

Youtube

A quick 20 min introduction to various UQ methods for Deep Learning:-

• Why is UQ required for Deep Learning
• Bayesian NN
• Monte Carlo Dropout
• MCMC
• Variational Inference
• Laplace Approximation
• Deep Ensembles
• Deep Evidence Regression