Challenges in interpretability of additive models

Xinyu Zhang; Julien Martinelli; Ti John

Challenges in interpretability of additive models

Xinyu Zhang, Julien Martinelli, Ti John

Université de Bordeaux

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Konferenssiesitys › Scientific › vertaisarvioitu

41 Lataukset (Pure)

Abstrakti

We review generalized additive models as a type of ‘transparent’ model that has recently seen renewed interest in the deep learning community as neural additive models. We highlight multiple types of nonidentifiability in this model class and discuss challenges in interpretability, arguing for restraint when claiming ‘interpretability’ or ‘suitability for safety-critical applications’ of such models.

Alkuperäiskieli	Englanti
Tila	Julkaistu - 2024
OKM-julkaisutyyppi	Ei sovellu
Tapahtuma	Workshop on Explainable Artificial Intelligence - Virtual, Online Kesto: 14 elok. 2024 → 15 elok. 2024 https://sites.google.com/view/xai2024/

Workshop

Workshop	Workshop on Explainable Artificial Intelligence
Lyhennettä	XAI
Kaupunki	Virtual, Online
Ajanjakso	14/08/2024 → 15/08/2024
www-osoite	https://sites.google.com/view/xai2024/

Pääsy asiakirjaan

Zhang_xai2024Final published version, 1,97 MB

https://drive.google.com/file/d/1ap0WJGnUMGz1GAZaqPQsjSq0RgZZWB9v/view?usp=sharing

1 Aktiivinen
2 Päättynyt

HEALED/Kaski S.: Human-steered next-generation machine learning for reviving drug design (HEALED)
Kaski, S. (Vastuullinen tutkija), Martinelli, J. (Projektin jäsen), Naumov, A. (Projektin jäsen), Zhang, X. (Projektin jäsen), Stefanović, O. (Projektin jäsen), Barabanshchikova, P. (Projektin jäsen), Zhu, L. (Projektin jäsen), Nahal, Y. (Projektin jäsen) & Masood, A. (Projektin jäsen)
01/09/2021 → 31/08/2025
Projekti: Academy of Finland: Other research funding
ELISE: European Learning and Intelligent Systems Excellence
Kaski, S. (Vastuullinen tutkija)
01/09/2020 → 31/08/2024
Projekti: EU: Framework programmes funding
-: Finnish Center for Artificial Intelligence
Kaski, S. (Vastuullinen tutkija)
01/01/2019 → 31/12/2022
Projekti: Academy of Finland: Other research funding

Science-IT
Hakala, M. (Manager)
Perustieteiden korkeakoulu
Laitteistot/tilat: Facility

Siteeraa tätä

@conference{d676c717c80446aa9026674799e54105,

title = "Challenges in interpretability of additive models",

abstract = "We review generalized additive models as a type of {\textquoteleft}transparent{\textquoteright} model that has recently seen renewed interest in the deep learning community as neural additive models. We highlight multiple types of nonidentifiability in this model class and discuss challenges in interpretability, arguing for restraint when claiming {\textquoteleft}interpretability{\textquoteright} or {\textquoteleft}suitability for safety-critical applications{\textquoteright} of such models.",

author = "Xinyu Zhang and Julien Martinelli and Ti John",

note = "| openaire: EC/H2020/951847/EU//ELISE ; Workshop on Explainable Artificial Intelligence, XAI ; Conference date: 14-08-2024 Through 15-08-2024",

year = "2024",

language = "English",

url = "https://sites.google.com/view/xai2024/",

}

TY - CONF

T1 - Challenges in interpretability of additive models

AU - Zhang, Xinyu

AU - Martinelli, Julien

AU - John, Ti

N1 - | openaire: EC/H2020/951847/EU//ELISE

PY - 2024

Y1 - 2024

N2 - We review generalized additive models as a type of ‘transparent’ model that has recently seen renewed interest in the deep learning community as neural additive models. We highlight multiple types of nonidentifiability in this model class and discuss challenges in interpretability, arguing for restraint when claiming ‘interpretability’ or ‘suitability for safety-critical applications’ of such models.

AB - We review generalized additive models as a type of ‘transparent’ model that has recently seen renewed interest in the deep learning community as neural additive models. We highlight multiple types of nonidentifiability in this model class and discuss challenges in interpretability, arguing for restraint when claiming ‘interpretability’ or ‘suitability for safety-critical applications’ of such models.

M3 - Paper

T2 - Workshop on Explainable Artificial Intelligence

Y2 - 14 August 2024 through 15 August 2024

ER -

Challenges in interpretability of additive models

Abstrakti

Workshop

Pääsy asiakirjaan

Sormenjälki

Projektit

HEALED/Kaski S.: Human-steered next-generation machine learning for reviving drug design (HEALED)

ELISE: European Learning and Intelligent Systems Excellence

-: Finnish Center for Artificial Intelligence

Laitteet

Science-IT

Siteeraa tätä