Saved in:
Bibliographic Details
Main Authors: Sun, Yiyang, Chen, Zhi, Orlandi, Vittorio, Wang, Tong, Rudin, Cynthia
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2402.09702
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917609276964864
author Sun, Yiyang
Chen, Zhi
Orlandi, Vittorio
Wang, Tong
Rudin, Cynthia
author_facet Sun, Yiyang
Chen, Zhi
Orlandi, Vittorio
Wang, Tong
Rudin, Cynthia
contents Even if a model is not globally sparse, it is possible for decisions made from that model to be accurately and faithfully described by a small number of features. For instance, an application for a large loan might be denied to someone because they have no credit history, which overwhelms any evidence towards their creditworthiness. In this work, we introduce the Sparse Explanation Value (SEV), a new way of measuring sparsity in machine learning models. In the loan denial example above, the SEV is 1 because only one factor is needed to explain why the loan was denied. SEV is a measure of decision sparsity rather than overall model sparsity, and we are able to show that many machine learning models -- even if they are not sparse -- actually have low decision sparsity, as measured by SEV. SEV is defined using movements over a hypercube, allowing SEV to be defined consistently over various model classes, with movement restrictions reflecting real-world constraints. We proposed the algorithms that reduce SEV without sacrificing accuracy, providing sparse and completely faithful explanations, even without globally sparse models.
format Preprint
id arxiv_https___arxiv_org_abs_2402_09702
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Sparse and Faithful Explanations Without Sparse Models
Sun, Yiyang
Chen, Zhi
Orlandi, Vittorio
Wang, Tong
Rudin, Cynthia
Machine Learning
Even if a model is not globally sparse, it is possible for decisions made from that model to be accurately and faithfully described by a small number of features. For instance, an application for a large loan might be denied to someone because they have no credit history, which overwhelms any evidence towards their creditworthiness. In this work, we introduce the Sparse Explanation Value (SEV), a new way of measuring sparsity in machine learning models. In the loan denial example above, the SEV is 1 because only one factor is needed to explain why the loan was denied. SEV is a measure of decision sparsity rather than overall model sparsity, and we are able to show that many machine learning models -- even if they are not sparse -- actually have low decision sparsity, as measured by SEV. SEV is defined using movements over a hypercube, allowing SEV to be defined consistently over various model classes, with movement restrictions reflecting real-world constraints. We proposed the algorithms that reduce SEV without sacrificing accuracy, providing sparse and completely faithful explanations, even without globally sparse models.
title Sparse and Faithful Explanations Without Sparse Models
topic Machine Learning
url https://arxiv.org/abs/2402.09702