ciu.image: An R Package for Explaining Image Classification with Contextual Importance and Utility

Kary Främling; Samanta Knapic̆; Avleen Malhi

doi:10.1007/978-3-030-82017-6_4

ciu.image: An R Package for Explaining Image Classification with Contextual Importance and Utility

Kary Främling^*, Samanta Knapic̆, Avleen Malhi

^*Tämän työn vastaava kirjoittaja

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference contribution › Scientific › vertaisarvioitu

18 Lataukset (Pure)

Abstrakti

Many techniques have been proposed in recent years that attempt to explain results of image classifiers, notably for the case when the classifier is a deep neural network. This paper presents an implementation of the Contextual Importance and Utility method for explaining image classifications. It is an R package that can be used with the most usual image classification models. The paper shows results for typical benchmark images, as well as for a medical data set of gastro-enterological images. For comparison, results produced by the LIME method are included. Results show that CIU produces similar or better results than LIME with significantly shorter calculation times. However, the main purpose of this paper is to bring the existence of this package to general knowledge and use, rather than comparing with other explanation methods.

Alkuperäiskieli	Englanti
Otsikko	Explainable and Transparent AI and Multi-Agent Systems - 3rd International Workshop, EXTRAAMAS 2021, Revised Selected Papers
Toimittajat	Davide Calvaresi, Amro Najjar, Michael Winikoff, Kary Främling
Sivut	55-62
Sivumäärä	8
DOI - pysyväislinkit	https://doi.org/10.1007/978-3-030-82017-6_4
Tila	Julkaistu - 2021
OKM-julkaisutyyppi	A4 Artikkeli konferenssijulkaisuussa
Tapahtuma	International Workshop on Explainable, Transparent AI and Multi-Agent Systems - Virtual, Online Kesto: 3 toukok. 2021 → 7 toukok. 2021 Konferenssinumero: 3

Julkaisusarja

Nimi	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Kustantaja	Springer
Vuosikerta	12688 LNAI
ISSN (painettu)	0302-9743
ISSN (elektroninen)	1611-3349

Workshop

Workshop	International Workshop on Explainable, Transparent AI and Multi-Agent Systems
Lyhennettä	EXTRAAMAS
Kaupunki	Virtual, Online
Ajanjakso	03/05/2021 → 07/05/2021

Pääsy asiakirjaan

10.1007/978-3-030-82017-6_4

SCI_Främling_CIU_imageAccepted author manuscript, 2,29 MB

SFX-saatavuus

Koko teksti

Muut tiedostot ja linkit

Linkki julkaisuun Scopuksessa

Siteeraa tätä

Främling, K., Knapic̆, S., & Malhi, A. (2021). ciu.image: An R Package for Explaining Image Classification with Contextual Importance and Utility. teoksessa D. Calvaresi, A. Najjar, M. Winikoff, & K. Främling (Toimittajat), Explainable and Transparent AI and Multi-Agent Systems - 3rd International Workshop, EXTRAAMAS 2021, Revised Selected Papers (Sivut 55-62). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vuosikerta 12688 LNAI). https://doi.org/10.1007/978-3-030-82017-6_4

Främling, Kary ; Knapic̆, Samanta ; Malhi, Avleen. / ciu.image : An R Package for Explaining Image Classification with Contextual Importance and Utility. Explainable and Transparent AI and Multi-Agent Systems - 3rd International Workshop, EXTRAAMAS 2021, Revised Selected Papers. Toimittaja / Davide Calvaresi ; Amro Najjar ; Michael Winikoff ; Kary Främling. 2021. Sivut 55-62 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{b450569b555746b4b4058b3b76e798f6,

title = "ciu.image: An R Package for Explaining Image Classification with Contextual Importance and Utility",

abstract = "Many techniques have been proposed in recent years that attempt to explain results of image classifiers, notably for the case when the classifier is a deep neural network. This paper presents an implementation of the Contextual Importance and Utility method for explaining image classifications. It is an R package that can be used with the most usual image classification models. The paper shows results for typical benchmark images, as well as for a medical data set of gastro-enterological images. For comparison, results produced by the LIME method are included. Results show that CIU produces similar or better results than LIME with significantly shorter calculation times. However, the main purpose of this paper is to bring the existence of this package to general knowledge and use, rather than comparing with other explanation methods.",

keywords = "Contextual importance and utility, Deep neural network, Explainable artificial intelligence, Image classification",

author = "Kary Fr{\"a}mling and Samanta Knapi{\u c} and Avleen Malhi",

note = "Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; International Workshop on Explainable, Transparent AI and Multi-Agent Systems, EXTRAAMAS ; Conference date: 03-05-2021 Through 07-05-2021",

year = "2021",

doi = "10.1007/978-3-030-82017-6_4",

language = "English",

isbn = "9783030820169",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "55--62",

editor = "Davide Calvaresi and Amro Najjar and Michael Winikoff and Kary Fr{\"a}mling",

booktitle = "Explainable and Transparent AI and Multi-Agent Systems - 3rd International Workshop, EXTRAAMAS 2021, Revised Selected Papers",

}

Främling, K, Knapic̆, S & Malhi, A 2021, ciu.image: An R Package for Explaining Image Classification with Contextual Importance and Utility. julkaisussa D Calvaresi, A Najjar, M Winikoff & K Främling (toim), Explainable and Transparent AI and Multi-Agent Systems - 3rd International Workshop, EXTRAAMAS 2021, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vuosikerta 12688 LNAI, Sivut 55-62, International Workshop on Explainable, Transparent AI and Multi-Agent Systems, Virtual, Online, 03/05/2021. https://doi.org/10.1007/978-3-030-82017-6_4

ciu.image : An R Package for Explaining Image Classification with Contextual Importance and Utility. / Främling, Kary; Knapic̆, Samanta; Malhi, Avleen.

Explainable and Transparent AI and Multi-Agent Systems - 3rd International Workshop, EXTRAAMAS 2021, Revised Selected Papers. toim. / Davide Calvaresi; Amro Najjar; Michael Winikoff; Kary Främling. 2021. s. 55-62 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vuosikerta 12688 LNAI).

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference contribution › Scientific › vertaisarvioitu

TY - GEN

T1 - ciu.image

T2 - International Workshop on Explainable, Transparent AI and Multi-Agent Systems

AU - Främling, Kary

AU - Knapic̆, Samanta

AU - Malhi, Avleen

N1 - Conference code: 3

PY - 2021

Y1 - 2021

N2 - Many techniques have been proposed in recent years that attempt to explain results of image classifiers, notably for the case when the classifier is a deep neural network. This paper presents an implementation of the Contextual Importance and Utility method for explaining image classifications. It is an R package that can be used with the most usual image classification models. The paper shows results for typical benchmark images, as well as for a medical data set of gastro-enterological images. For comparison, results produced by the LIME method are included. Results show that CIU produces similar or better results than LIME with significantly shorter calculation times. However, the main purpose of this paper is to bring the existence of this package to general knowledge and use, rather than comparing with other explanation methods.

AB - Many techniques have been proposed in recent years that attempt to explain results of image classifiers, notably for the case when the classifier is a deep neural network. This paper presents an implementation of the Contextual Importance and Utility method for explaining image classifications. It is an R package that can be used with the most usual image classification models. The paper shows results for typical benchmark images, as well as for a medical data set of gastro-enterological images. For comparison, results produced by the LIME method are included. Results show that CIU produces similar or better results than LIME with significantly shorter calculation times. However, the main purpose of this paper is to bring the existence of this package to general knowledge and use, rather than comparing with other explanation methods.

KW - Contextual importance and utility

KW - Deep neural network

KW - Explainable artificial intelligence

KW - Image classification

UR - http://www.scopus.com/inward/record.url?scp=85113311249&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-82017-6_4

DO - 10.1007/978-3-030-82017-6_4

M3 - Conference contribution

AN - SCOPUS:85113311249

SN - 9783030820169

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 55

EP - 62

BT - Explainable and Transparent AI and Multi-Agent Systems - 3rd International Workshop, EXTRAAMAS 2021, Revised Selected Papers

A2 - Calvaresi, Davide

A2 - Najjar, Amro

A2 - Winikoff, Michael

A2 - Främling, Kary

Y2 - 3 May 2021 through 7 May 2021

ER -

Främling K, Knapic̆ S, Malhi A. ciu.image: An R Package for Explaining Image Classification with Contextual Importance and Utility. julkaisussa Calvaresi D, Najjar A, Winikoff M, Främling K, toimittajat, Explainable and Transparent AI and Multi-Agent Systems - 3rd International Workshop, EXTRAAMAS 2021, Revised Selected Papers. 2021. s. 55-62. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-82017-6_4

ciu.image: An R Package for Explaining Image Classification with Contextual Importance and Utility

Abstrakti

Julkaisusarja

Workshop

Pääsy asiakirjaan

SFX-saatavuus

Muut tiedostot ja linkit

Sormenjälki

Siteeraa tätä