Probabilistic Deep Learning with Probabilistic Neural Networks and Deep Probabilistic Models

By Daniel T. Chang et al

Read the original document by opening this link in a new tab.

Table of Contents

1 Introduction
2 Background Information
3 Probabilistic Neural Networks
4 Deep Probabilistic Models

Summary

Probabilistic deep learning is based on the use of probabilistic models and deep neural networks, accounting for uncertainty. It distinguishes between probabilistic neural networks and deep probabilistic models. TensorFlow Probability is a library for probabilistic modeling and inference. Uncertainty in deep learning includes model uncertainty and data uncertainty. Probabilistic models use probability theory to represent uncertainty and for inference. Bayesian inference is used for learning from data. TensorFlow Probability provides probabilistic building blocks for modeling and inference. Probabilistic neural networks utilize probabilistic layers to represent uncertainty. Bayesian neural networks and mixture density networks are examples. Deep probabilistic models incorporate deep neural network components to capture complex relationships between random variables. Variational autoencoders, deep Gaussian processes, and deep mixed effects models are examples. TensorFlow Probability provides tools for building both types of models.