Robust Explainability: A tutorial on gradient-based attribution methods for deep neural networks

Ian E. Nielsen; Dimah Dera; Ghulam Rasool; Ravi P. Ramachandran; Nidhal Carla Bouaynaya

doi:10.1109/MSP.2022.3142719

Robust Explainability: A tutorial on gradient-based attribution methods for deep neural networks

Ian E. Nielsen, Dimah Dera, Ghulam Rasool, Ravi P. Ramachandran, Nidhal Carla Bouaynaya

Electrical Engineering

Research output: Contribution to journal › Article › peer-review

72 Scopus citations

Abstract

The rise in deep neural networks (DNNs) has led to increased interest in explaining their predictions. While many methods for this exist, there is currently no consensus on how to evaluate them. On the other hand, robustness is a popular topic for deep learning (DL) research; however, it has been hardly talked about in explainability until very recently.

Original language	English (US)
Pages (from-to)	73-84
Number of pages	12
Journal	IEEE Signal Processing Magazine
Volume	39
Issue number	4
DOIs	https://doi.org/10.1109/MSP.2022.3142719
State	Published - Jul 1 2022

All Science Journal Classification (ASJC) codes

Signal Processing
Electrical and Electronic Engineering
Applied Mathematics

Access to Document

10.1109/MSP.2022.3142719

Cite this

@article{c72dc94073f44ffe94ce825ce0ed5d2a,

title = "Robust Explainability: A tutorial on gradient-based attribution methods for deep neural networks",

abstract = "The rise in deep neural networks (DNNs) has led to increased interest in explaining their predictions. While many methods for this exist, there is currently no consensus on how to evaluate them. On the other hand, robustness is a popular topic for deep learning (DL) research; however, it has been hardly talked about in explainability until very recently.",

author = "Nielsen, {Ian E.} and Dimah Dera and Ghulam Rasool and Ramachandran, {Ravi P.} and Bouaynaya, {Nidhal Carla}",

note = "Publisher Copyright: {\textcopyright} 1991-2012 IEEE.",

year = "2022",

month = jul,

day = "1",

doi = "10.1109/MSP.2022.3142719",

language = "English (US)",

volume = "39",

pages = "73--84",

journal = "IEEE Signal Processing Magazine",

issn = "1053-5888",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "4",

}

TY - JOUR

T1 - Robust Explainability

T2 - A tutorial on gradient-based attribution methods for deep neural networks

AU - Nielsen, Ian E.

AU - Dera, Dimah

AU - Rasool, Ghulam

AU - Ramachandran, Ravi P.

AU - Bouaynaya, Nidhal Carla

PY - 2022/7/1

Y1 - 2022/7/1

N2 - The rise in deep neural networks (DNNs) has led to increased interest in explaining their predictions. While many methods for this exist, there is currently no consensus on how to evaluate them. On the other hand, robustness is a popular topic for deep learning (DL) research; however, it has been hardly talked about in explainability until very recently.

AB - The rise in deep neural networks (DNNs) has led to increased interest in explaining their predictions. While many methods for this exist, there is currently no consensus on how to evaluate them. On the other hand, robustness is a popular topic for deep learning (DL) research; however, it has been hardly talked about in explainability until very recently.

UR - http://www.scopus.com/inward/record.url?scp=85133762306&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85133762306&partnerID=8YFLogxK

U2 - 10.1109/MSP.2022.3142719

DO - 10.1109/MSP.2022.3142719

M3 - Article

AN - SCOPUS:85133762306

SN - 1053-5888

VL - 39

SP - 73

EP - 84

JO - IEEE Signal Processing Magazine

JF - IEEE Signal Processing Magazine

IS - 4

ER -

Robust Explainability: A tutorial on gradient-based attribution methods for deep neural networks

Abstract

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this