The theory behind Bayesian active learning¶
In this document, we keep a list of the papers to get you started in Bayesian deep learning and Bayesian active learning.
We hope to include a summary for each of then in the future, but for now we have this list with some notes.
How to estimate uncertainty in Deep Learning networks¶
This is inspired by his publication Bayesian Deep Learning and a Probabilistic Perspective of Generalization
Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning (Gal and Ghahramani, 2016)
This describes Monte-Carlo Dropout, a way to estimate uncertainty through stochastic dropout at test time
Bayesian Uncertainty Estimation for Batch Normalized Deep Networks (Teye et al. 2018)
This describes Monte-Carlo BatchNorm, a way to estimate uncertainty through random batch norm parameters at test time
Bayesian Deep Learning and a Probabilistic Perspective of Generalization (Gordon Wilson and Izmailov, 2020)
Presentation of multi-SWAG a mix between VI and Ensembles.
Advances in Variational inference (Zhang et al, 2018)
Gives a quick introduction to VI and the most recent advances.
A Simple Baseline for Bayesian Uncertainty in Deep Learning (Maddox et al. 2019)
Presents SWAG, an easy way to create ensembles.
Bayesian active learning¶
Deep Bayesian Active Learning with Image Data (Gal and Islam and Ghahramani, 2017)
Fundamental paper on how to do Bayesian active learning. A must read.
Sampling bias in active learning (Dasgupta 2009)
Presents sampling bias and how to solve it by combining heuristics and random selection.
Bayesian Active Learning for Classification and Preference Learning (Houlsby et al. 2011)
Fundamental paper on one of the main heuristic BALD.
Bayesian active learning on NLP¶
Deep Bayesian Active Learning for Natural Language Processing: Results of a Large-Scale Empirical Study (Siddhant and Lipton, 2018)
Experimental paper on how to use Bayesian active learning on NLP tasks.