Predicting Uncertainty with Bayesian Neural Networks on MNIST Dataset

Least Confident Predictions


The objective of the project is class predictions with uncertainty with an implementation of a Bayesian Neural Network, trained and tested on the Rotated MNIST and Fashion MNIST datasets. The training process minimizes a loss function that considers both a Cross-Entropy loss and a Kullback-Leibler divergence loss term.

Chenhao Li
Chenhao Li
Reinforcement Learning for Robotics

My research interests focus on the general field of robot learning, including reinforcement learning, developmental robotics and legged intelligence.