Transférabilité - Historique des versions

Pitpitt : Remplacement de texte : « ↵ » par «  ==Sources== »

2024-01-28T18:53:24Z

Remplacement de texte : « ↵<small> » par «  ==Sources== »

← Version précédente		Version du 28 janvier 2024 à 14:53
Ligne 9 :		Ligne 9 :


	~~<small>~~		==Sources==

Pitpitt le 22 février 2023 à 13:36

2023-02-22T13:36:25Z

← Version précédente		Version du 22 février 2023 à 09:36
Ligne 15 :		Ligne 15 :


	[[Catégorie:~~publication~~]]		[[Catégorie:GRAND LEXIQUE FRANÇAIS]]

Imeziani le 22 février 2023 à 12:57

2023-02-22T12:57:26Z

← Version précédente		Version du 22 février 2023 à 08:57
Ligne 1 :		Ligne 1 :
	~~==en construction==~~

	== Définition ==		== Définition ==
	~~XXXXXXXXX~~		Capacité à acquérir et à réutiliser des connaissances en apprentissage profond. Elle est à l'origine de la quête à long terme visant à rendre l'apprentissage profond aussi efficace en termes de données que l'apprentissage humain, et a motivé la conception fructueuse d'algorithmes d'apprentissage profond plus puissants.

	== Français ==		== Français ==
	''' ~~XXXXXXXXX~~ '''		''' transférabilité '''

	== Anglais ==		== Anglais ==
	''' ~~Transferability~~'''		''' transferability'''

	One of the benefits of explainable ML models is the ability to understand how model weights can adapt to changes in the training environment which is essential to generalize across different scenarios.

	The success of deep learning algorithms generally depends on large-scale data, while humans appear to have inherent ability of knowledge transfer, by recognizing and applying relevant knowledge from previous learning experiences when encountering and solving unseen tasks. Such an ability to acquire and reuse knowledge is known as transferability in deep learning. It has formed the long-term quest towards making deep learning as data-efficient as human learning, and has been motivating fruitful design of more powerful deep learning algorithms. We present this survey to connect different isolated areas in deep learning with their relation to transferability, and to provide a unified and complete view to investigating transferability through the whole lifecycle of deep learning. The survey elaborates the fundamental goals and challenges in parallel with the core principles and methods, covering recent cornerstones in deep architectures, pre-training, task adaptation and domain adaptation. This highlights unanswered questions on the appropriate objectives for learning transferable knowledge and for adapting the knowledge to new tasks and domains, avoiding catastrophic forgetting and negative transfer. Finally, we implement a benchmark and an open-source library, enabling a fair evaluation of deep learning methods in terms of transferability.

	<small>		<small>
Ligne 20 :		Ligne 15 :


	[[Catégorie:~~vocabulary~~]]		[[Catégorie:publication]]

Imeziani : Imeziani a déplacé la page Transferability vers Transférabilité

2023-02-22T12:55:19Z

Imeziani a déplacé la page Transferability vers Transférabilité

← Version précédente	Version du 22 février 2023 à 08:55
(Aucune différence)

Pitpitt : Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' XXXXXXXXX ''' == Anglais == ''' Transferability''' One of the benefits of explainable ML models is... »

2023-02-07T13:12:25Z

Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' XXXXXXXXX ''' == Anglais == ''' Transferability''' One of the benefits of explainable ML models is... »

Nouvelle page

==en construction==

== Définition ==
XXXXXXXXX

== Français ==
''' XXXXXXXXX '''

== Anglais ==
''' Transferability'''

One of the benefits of explainable ML models is the ability to understand how model weights can adapt to changes in the training environment which is essential to generalize across different scenarios.

The success of deep learning algorithms generally depends on large-scale data, while humans appear to have inherent ability of knowledge transfer, by recognizing and applying relevant knowledge from previous learning experiences when encountering and solving unseen tasks. Such an ability to acquire and reuse knowledge is known as transferability in deep learning. It has formed the long-term quest towards making deep learning as data-efficient as human learning, and has been motivating fruitful design of more powerful deep learning algorithms. We present this survey to connect different isolated areas in deep learning with their relation to transferability, and to provide a unified and complete view to investigating transferability through the whole lifecycle of deep learning. The survey elaborates the fundamental goals and challenges in parallel with the core principles and methods, covering recent cornerstones in deep architectures, pre-training, task adaptation and domain adaptation. This highlights unanswered questions on the appropriate objectives for learning transferable knowledge and for adapting the knowledge to new tasks and domains, avoiding catastrophic forgetting and negative transfer. Finally, we implement a benchmark and an open-source library, enabling a fair evaluation of deep learning methods in terms of transferability.

<small>

[https://arxiv.org/abs/2201.05867 Source : arxiv]

[[Catégorie:vocabulary]]

Transférabilité - Historique des versions

Pitpitt : Remplacement de texte : « ↵ » par « ==Sources== »

Pitpitt le 22 février 2023 à 13:36

Imeziani le 22 février 2023 à 12:57

Imeziani : Imeziani a déplacé la page Transferability vers Transférabilité

Pitpitt : Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' XXXXXXXXX ''' == Anglais == ''' Transferability''' One of the benefits of explainable ML models is... »

Pitpitt : Remplacement de texte : « ↵ » par «  ==Sources== »

Imeziani : Imeziani a déplacé la page Transferability vers Transférabilité

Pitpitt : Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' XXXXXXXXX ''' == Anglais == ''' Transferability''' One of the benefits of explainable ML models is... »