Zobrazeno 1 - 10
of 3 559
pro vyhledávání: '"Boutry, A."'
Challenges persist in providing interpretable explanations for neural network reasoning in explainable AI (xAI). Existing methods like Integrated Gradients produce noisy maps, and LIME, while intuitive, may deviate from the model's reasoning. We intr
Externí odkaz:
http://arxiv.org/abs/2406.13257
Autor:
Doh, Miriam, Rodrigues, Caroline Mazini, Boutry, Nicolas, Najman, Laurent, Mancas, Matei, Bersini, Hugues
With Artificial Intelligence (AI) influencing the decision-making process of sensitive applications such as Face Verification, it is fundamental to ensure the transparency, fairness, and accountability of decisions. Although Explainable Artificial In
Externí odkaz:
http://arxiv.org/abs/2403.08789
The explication of Convolutional Neural Networks (CNN) through xAI techniques often poses challenges in interpretation. The inherent complexity of input features, notably pixels extracted from images, engenders complex correlations. Gradient-based me
Externí odkaz:
http://arxiv.org/abs/2401.14434
Publikováno v:
EPTCS 398, 2024, pp. 73-84
In 1926-1927, Tarski designed a set of axioms for Euclidean geometry which reached its final form in a manuscript by Schwabh\"auser, Szmielew and Tarski in 1983. The differences amount to simplifications obtained by Tarski and Gupta. Gupta presented
Externí odkaz:
http://arxiv.org/abs/2401.11904
Publikováno v:
Sojourn: Journal of Social Issues in Southeast Asia, 2024 Mar 01. 39(1), 28-37.
Externí odkaz:
https://www.jstor.org/stable/27300973
Autor:
Boutry, Maxime, Ivanoff, Jacques
Publikováno v:
Sojourn: Journal of Social Issues in Southeast Asia, 2024 Mar 01. 39(1), 154-180.
Externí odkaz:
https://www.jstor.org/stable/27300978
Autor:
Sophie Tissot, Jordan Meliani, Matthew Chee, Aurora M. Nedelcu, Justine Boutry, Jácint Tökölyi, Rodrigo Hamede, Benjamin Roche, Beata Ujvari, Frédéric Thomas, Antoine M. Dujon
Publikováno v:
Scientific Reports, Vol 14, Iss 1, Pp 1-7 (2024)
Abstract Recent theoretical advances in the One Health approach have suggested that cancer pathologies should be given greater consideration, as cancers often render their hosts more vulnerable to infectious agents, which could turn them into super s
Externí odkaz:
https://doaj.org/article/586d2919e7bf4d62a3eca1ddbddb6b55
Providing interpretability of deep-learning models to non-experts, while fundamental for a responsible real-world usage, is challenging. Attribution maps from xAI techniques, such as Integrated Gradients, are a typical example of a visualization tech
Externí odkaz:
http://arxiv.org/abs/2309.00018
Autor:
Sarah Meulebrouck, Judith Merrheim, Gurvan Queniat, Cyril Bourouh, Mehdi Derhourhi, Mathilde Boissel, Xiaoyan Yi, Alaa Badreddine, Raphaël Boutry, Audrey Leloire, Bénédicte Toussaint, Souhila Amanzougarene, Emmanuel Vaillant, Emmanuelle Durand, Hélène Loiselle, Marlène Huyvaert, Aurélie Dechaume, Victoria Scherrer, Piero Marchetti, Beverley Balkau, Guillaume Charpentier, Sylvia Franc, Michel Marre, Ronan Roussel, Raphaël Scharfmann, Miriam Cnop, Mickaël Canouil, Morgane Baron, Philippe Froguel, Amélie Bonnefond
Publikováno v:
Nature Communications, Vol 15, Iss 1, Pp 1-12 (2024)
Abstract Functional genetics has identified drug targets for metabolic disorders. Opioid use impacts metabolic homeostasis, although mechanisms remain elusive. Here, we explore the OPRD1 gene (encoding delta opioid receptor, DOP) to understand its im
Externí odkaz:
https://doaj.org/article/78b374b3ff9b43728d9fb6ee8aab174e
Any watershed, when defined on a stack on a normal pseudomanifold of dimension d, is a pure (d -- 1)-subcomplex that satisfies a drop-of-water principle. In this paper, we introduce Morse stacks, a class of functions that are equivalent to discrete M
Externí odkaz:
http://arxiv.org/abs/2301.03840