Toward Interpretable Topic Discovery via Anchored Correlation Explanation.
K. Reing, D. Kale, G. Steeg, and A. Galstyan. (2016)cite arxiv:1606.07043Comment: presented at 2016 ICML Workshop on #Data4Good: Machine Learning in Social Good Applications, New York, NY.
Abstract
Many predictive tasks, such as diagnosing a patient based on their medical
chart, are ultimately defined by the decisions of human experts. Unfortunately,
encoding experts' knowledge is often time consuming and expensive. We propose a
simple way to use fuzzy and informal knowledge from experts to guide discovery
of interpretable latent topics in text. The underlying intuition of our
approach is that latent factors should be informative about both correlations
in the data and a set of relevance variables specified by an expert.
Mathematically, this approach is a combination of the information bottleneck
and Total Correlation Explanation (CorEx). We give a preliminary evaluation of
Anchored CorEx, showing that it produces more coherent and interpretable topics
on two distinct corpora.
Description
[1606.07043] Toward Interpretable Topic Discovery via Anchored Correlation Explanation
%0 Generic
%1 reing2016toward
%A Reing, Kyle
%A Kale, David C.
%A Steeg, Greg Ver
%A Galstyan, Aram
%D 2016
%K classification corex fulltext machinelearning
%T Toward Interpretable Topic Discovery via Anchored Correlation Explanation.
%U http://arxiv.org/abs/1606.07043
%X Many predictive tasks, such as diagnosing a patient based on their medical
chart, are ultimately defined by the decisions of human experts. Unfortunately,
encoding experts' knowledge is often time consuming and expensive. We propose a
simple way to use fuzzy and informal knowledge from experts to guide discovery
of interpretable latent topics in text. The underlying intuition of our
approach is that latent factors should be informative about both correlations
in the data and a set of relevance variables specified by an expert.
Mathematically, this approach is a combination of the information bottleneck
and Total Correlation Explanation (CorEx). We give a preliminary evaluation of
Anchored CorEx, showing that it produces more coherent and interpretable topics
on two distinct corpora.
@misc{reing2016toward,
abstract = {Many predictive tasks, such as diagnosing a patient based on their medical
chart, are ultimately defined by the decisions of human experts. Unfortunately,
encoding experts' knowledge is often time consuming and expensive. We propose a
simple way to use fuzzy and informal knowledge from experts to guide discovery
of interpretable latent topics in text. The underlying intuition of our
approach is that latent factors should be informative about both correlations
in the data and a set of relevance variables specified by an expert.
Mathematically, this approach is a combination of the information bottleneck
and Total Correlation Explanation (CorEx). We give a preliminary evaluation of
Anchored CorEx, showing that it produces more coherent and interpretable topics
on two distinct corpora.},
added-at = {2017-02-18T20:04:19.000+0100},
author = {Reing, Kyle and Kale, David C. and Steeg, Greg Ver and Galstyan, Aram},
biburl = {https://www.bibsonomy.org/bibtex/21633b6b9d5e8b084c6c4e59fbb35a89f/marcsaric},
description = {[1606.07043] Toward Interpretable Topic Discovery via Anchored Correlation Explanation},
interhash = {a0ff0c0ff14a254a77f4c593c8a9377b},
intrahash = {1633b6b9d5e8b084c6c4e59fbb35a89f},
keywords = {classification corex fulltext machinelearning},
note = {cite arxiv:1606.07043Comment: presented at 2016 ICML Workshop on #Data4Good: Machine Learning in Social Good Applications, New York, NY},
timestamp = {2017-05-05T22:14:49.000+0200},
title = {Toward Interpretable Topic Discovery via Anchored Correlation Explanation.},
url = {http://arxiv.org/abs/1606.07043},
year = 2016
}