Uncertainty in the Variational Information Bottleneck

07/02/2018
by   Alexander A. Alemi, et al.
2

We present a simple case study, demonstrating that Variational Information Bottleneck (VIB) can improve a network's classification calibration as well as its ability to detect out-of-distribution data. Without explicitly being designed to do so, VIB gives two natural metrics for handling and quantifying uncertainty.

READ FULL TEXT

page 4

page 8

page 9

page 10

research
01/18/2021

Quantifying Uncertainty in Infectious Disease Mechanistic Models

This primer describes the statistical uncertainty in mechanistic models ...
research
06/08/2020

The Dual Information Bottleneck

The Information Bottleneck (IB) framework is a general characterization ...
research
10/27/2018

The Variational Deficiency Bottleneck

We introduce a bottleneck method for learning data representations based...
research
02/07/2023

IB-UQ: Information bottleneck based uncertainty quantification for neural function regression and neural operator learning

In this paper, a novel framework is established for uncertainty quantifi...
research
09/20/2019

Characterizing Sources of Uncertainty to Proxy Calibration and Disambiguate Annotator and Data Bias

Supporting model interpretability for complex phenomena where annotators...
research
05/24/2016

Relevant sparse codes with variational information bottleneck

In many applications, it is desirable to extract only the relevant aspec...
research
01/10/2013

Multivariate Information Bottleneck

The Information bottleneck method is an unsupervised non-parametric data...

Please sign up or login with your details

Forgot password? Click here to reset