DataFranca - Modifications récentes [fr]

Planning

Pitpitt — Mon, 20 Apr 2026 01:18:02 GMT

Page redirigée vers Planification

Nouvelle page

#REDIRECTION[[Planification]]

Data visualization

Pitpitt — Mon, 20 Apr 2026 01:16:56 GMT

Destination de redirection modifiée de Visualisation de données en Visualisation des données

← Version précédente		Version du 19 avril 2026 à 21:16
Ligne 1 :		Ligne 1 :
	#REDIRECTION[[Visualisation de données]]		#REDIRECTION[[Visualisation des données]]

	[[Catégorie:ENGLISH]]		[[Catégorie:ENGLISH]]

Visualisation des données

Pitpitt — Mon, 20 Apr 2026 01:15:57 GMT

← Version précédente		Version du 19 avril 2026 à 21:15
Ligne 20 :		Ligne 20 :

	'''big data visualization'''		'''big data visualization'''

			Data visualization is the creative use of raw data, statistics and graphic techniques to produce graphical representations capable of making data more easily interpretable or highlighting trends. Data visualization aids decision-making within an organization.

			There are three main categories of graphic representations used in data visualization: computer graphics (e.g., the age pyramid of a population at a given point in time), animated visualizations (e.g., the evolution of life expectancy over the centuries) and interactive visualizations (e.g., the result of an election according to postal code).


	==Español==		==Español==

Vision artificielle

Pitpitt — Mon, 20 Apr 2026 01:15:25 GMT

← Version précédente		Version du 19 avril 2026 à 21:15
Ligne 17 :		Ligne 17 :

	'''artificial vision''' <small>(<i>usage plus rare</i>)</small>		'''artificial vision''' <small>(<i>usage plus rare</i>)</small>

			Computer vision is the art and science of enabling computers to interpret images intelligently. Using visual data, a computer vision system can learn to accurately identify, analyze and classify objects, and make decisions based on what it sees.

	==Español==		==Español==

Vecteur-mot

Pitpitt — Mon, 20 Apr 2026 01:14:56 GMT

← Version précédente		Version du 19 avril 2026 à 21:14
Ligne 32 :		Ligne 32 :

	'''word vector'''		'''word vector'''

			A word vector is a distributed, dense representation of a word in the form of real numbers, or an embedding. Although often applied to words, word embedding modeling is not limited to words, and can be applied to sentences, paragraphs, documents and more.

			Word embedding representation has the property that words appearing in similar contexts have vectors that are relatively close. As a result, the distance between word vectors of similar concepts, such as bottle and cup, is smaller than that between semantically distant concepts, such as bottle and gnat.

			A word embedding is obtained by machine learning or by analyzing co-occurrences or data neighborhoods.


	==Español==		==Español==

Traitement automatique de la langue naturelle

Pitpitt — Mon, 20 Apr 2026 01:13:33 GMT

← Version précédente		Version du 19 avril 2026 à 21:13
Ligne 46 :		Ligne 46 :

	'''computational linguistics'''		'''computational linguistics'''

			The natural language processing (NLP) is a multidisciplinary field involving linguistics, computer science and artificial intelligence.
			The aim of NLP is to study and understand natural language, and to create tools for various applications.
			NLP is gradually moving out of research laboratories and into computer applications requiring the integration of human language with machines. It is currently used in :
			* grammar correction,
			* search engines
			* chatbot,
			* text generation,
			* automatic transcription,
			* speech synthesis,
			* speech recognition,
			* machine translation,
			* text summarization, etc.


	==Español==		==Español==

Traduction automatique

Pitpitt — Mon, 20 Apr 2026 01:11:22 GMT

← Version précédente		Version du 19 avril 2026 à 21:11
Ligne 18 :		Ligne 18 :

	'''automatic translation'''		'''automatic translation'''

			Machine translation is a sub-field of artificial intelligence, or more precisely natural language processing, which automatically translates text from one source language into another, the target language, without human intervention.

			Université de Montréal is a pioneering institution in machine translation research, from TAUM's work in the 1970s on translating weather reports to the recent advances in neural machine translation (NMT) by the MILA laboratory under the direction of Yoshua Bengio.


	==Español==		==Español==

Théorie des jeux

Pitpitt — Mon, 20 Apr 2026 01:10:54 GMT

← Version précédente		Version du 19 avril 2026 à 21:10
Ligne 9 :		Ligne 9 :
	==Anglais==		==Anglais==
	'''game theory'''		'''game theory'''

			The game theory is the study of mathematical models of strategic interaction between rational decision-makers. It has applications in all areas of the social sciences, as well as in logic and computer science.

			Originally, these were zero-sum games, in which one person's gains resulted in losses for the other participants.

			Today, game theory is applied to a wide range of behavioral relationships and is now a generic term for the science of logical decision-making in humans, animals and computers.


	==Español==		==Español==

Théorie des graphes

Pitpitt — Mon, 20 Apr 2026 01:10:28 GMT

← Version précédente		Version du 19 avril 2026 à 21:10
Ligne 9 :		Ligne 9 :
	==Anglais==		==Anglais==
	'''graph theory'''		'''graph theory'''

			The graph theory is the mathematical and computer science discipline that studies graphs as mathematical structures used to model relationships between objects in pairs.

			A graph is made up of vertices that are connected by edges, called nodes or points and lines or links respectively.


	==Español==		==Español==

Test de Turing

Pitpitt — Mon, 20 Apr 2026 01:04:22 GMT

← Version précédente		Version du 19 avril 2026 à 21:04
Ligne 20 :		Ligne 20 :

	'''imitation game '''		'''imitation game '''

			The Turing Test or imitation game was proposed by British mathematician Alan Turing in 1950 to test a computer's ability to impersonate a human in a natural language conversation.

			The test takes place between a human judge who initiates the conversation, another human and a computer or software that responds anonymously. Through textual exchanges, the judge must determine which of his or her interlocutors is the computer. The test therefore consists of putting a human in a blind verbal confrontation with a computer and another human. If the judge who initiates the conversations is unable to determine which of his or her interlocutors is a computer, the computer can be considered to have passed the Turing test.


	==Español==		==Español==

Système expert

Pitpitt — Mon, 20 Apr 2026 01:03:49 GMT

← Version précédente		Version du 19 avril 2026 à 21:03
Ligne 9 :		Ligne 9 :

	'''ES'''		'''ES'''

			A tool capable of reproducing the cognitive mechanisms of an expert in a particular field. This is one of the paths leading to artificial intelligence. More precisely, an expert system is software capable of answering questions by reasoning on the basis of known facts and rules.

			In particular, it can be used as a decision-making tool.

			The first expert system was Dendral. It was used to identify chemical constituents.


	==Español==		==Español==

Recommendation System

Pitpitt — Mon, 20 Apr 2026 01:02:04 GMT

Destination de redirection modifiée de Système de recommandations en Système de recommandation

@@ Ligne 1 : / Ligne 1 : @@
-#REDIRECTION[[Système de recommandations]]
+#REDIRECTION[[Système de recommandation]]
 [[Catégorie:ENGLISH]]
 [[Catégorie:Intelligence artificielle‏‎]]

Système de recommandation

Pitpitt — Mon, 20 Apr 2026 01:01:41 GMT

← Version précédente		Version du 19 avril 2026 à 21:01
Ligne 34 :		Ligne 34 :

	'''recommender platform'''		'''recommender platform'''

			A recommendation system offers the user assistance in searching for products and services, based on data and machine learning.

			More specifically, a recommendation system relies on a set of information filtering techniques that predict the rank or preference a user assigns to an item among a set of similar items (films, music, books, news, images, web pages, etc.) that are likely to interest them.

			Typically, a recommendation system compares a user's profile to their past choices or preferences, to users with similar profiles (collaborative filtering) or who have shown an interest in similar items (content-based filtering).

	==Español==		==Español==

Système d'IA autonome

Pitpitt — Mon, 20 Apr 2026 01:00:49 GMT

← Version précédente		Version du 19 avril 2026 à 21:00
Ligne 24 :		Ligne 24 :

	'''AS'''		'''AS'''

			In artificial intelligence, a system is said to be “autonomous” if, from the moment it is activated, it operates on its own, without human intervention.

			See also: intelligent agent


	==Español==		==Español==

Autonomous AI system

Pitpitt — Mon, 20 Apr 2026 00:59:13 GMT

Page redirigée vers Système d'IA autonome

Nouvelle page

#REDIRECTION[[Système d'IA autonome]]

Système à base de connaissances

Pitpitt — Mon, 20 Apr 2026 00:58:26 GMT

← Version précédente		Version du 19 avril 2026 à 20:58
Ligne 18 :		Ligne 18 :

	'''KBS'''		'''KBS'''

			Derived from symbolic artificial intelligence technologies, a knowledge-based system is software that uses a knowledge base to solve complex problems.

			The term “knowledge-based system” refers to many types of system, all of which are based on an explicit representation of knowledge, either a knowledge base and reasoning system, or an inference engine enabling it to derive new knowledge.


	==Español==		==Español==

Synthographie

Pitpitt — Mon, 20 Apr 2026 00:55:45 GMT

← Version précédente		Version du 19 avril 2026 à 20:55
Ligne 24 :		Ligne 24 :

	'''synthography'''		'''synthography'''

			The image creation by artificial intelligence for artistic purposes.

	==Español==		==Español==

Artificial intelligence art

Pitpitt — Mon, 20 Apr 2026 00:54:59 GMT

Page redirigée vers Synthographie

Nouvelle page

#REDIRECTION[[Synthographie]]

Surapprentissage

Pitpitt — Mon, 20 Apr 2026 00:54:15 GMT

← Version précédente		Version du 19 avril 2026 à 20:54
Ligne 18 :		Ligne 18 :

	'''overtraining'''		'''overtraining'''

			The overfitting is a problem that can arise in mathematical classification methods for neural networks. It is generally caused by incorrectly dimensioning the structure used to classify. Because of its excessive capacity to store information, a structure in an overlearning situation will have difficulty generalizing the features of the data. It then behaves like a table containing all the samples used during learning and loses its predictive powers on new samples.

			See also: curse of dimensionality

	==Español==		==Español==

Super-résolution

Pitpitt — Mon, 20 Apr 2026 00:53:41 GMT

← Version précédente		Version du 19 avril 2026 à 20:53
Ligne 37 :		Ligne 37 :

	'''SRGAN'''		'''SRGAN'''

			In computer vision and image processing, super-resolution refers to the process of improving the resolution, i.e. the level of detail, of an image.

	==Español==		==Español==

Sous-ajustement

Pitpitt — Mon, 20 Apr 2026 00:53:21 GMT

← Version précédente		Version du 19 avril 2026 à 20:53
Ligne 17 :		Ligne 17 :

	'''underlearning'''		'''underlearning'''

			A situation observed when a machine learning algorithm or statistical model only roughly fits the training data, resulting in a high error on the training data

	==Español==		==Español==

Singularité technologique

Pitpitt — Mon, 20 Apr 2026 00:52:29 GMT

← Version précédente		Version du 19 avril 2026 à 20:52
Ligne 16 :		Ligne 16 :

	'''singularity'''		'''singularity'''

			The technological singularity, or simply the singularity, is the hypothesis that the advent of strong artificial intelligence would trigger a surge in technological growth that would induce unpredictable changes in human society.

			Beyond this point, progress would only be achieved by AIs whose intelligence would increase. New generations of increasingly intelligent AIs would appear at an ever-faster pace, creating an “intelligence explosion” and, by extension, a powerful superintelligence that would far surpass human intelligence.

	==Español==		==Español==

Simulation IA

Pitpitt — Mon, 20 Apr 2026 00:50:44 GMT

← Version précédente		Version du 19 avril 2026 à 20:50
Ligne 9 :		Ligne 9 :
	== Anglais ==		== Anglais ==
	'''AI simulation '''		'''AI simulation '''

			AI simulation accelerates the discovery process by using AI to identify the most promising simulations to run on a massive data set. Equally important, it determines the computing infrastructure best suited to the task, be it a simple calculator or even a quantum computer

	==Español==		==Español==

Séparateur à vaste marge

Pitpitt — Mon, 20 Apr 2026 00:50:23 GMT

← Version précédente		Version du 19 avril 2026 à 20:50
Ligne 34 :		Ligne 34 :

	'''kernel method'''		'''kernel method'''

			The support vector machine (SVM) is a supervised learning algorithm based on a generalization of linear classifiers and designed to solve classification and regression problems. Created in the mid-90s by Vladimir Vapnik, SVMs have been applied to a wide range of fields, including bioinformatics, information retrieval, computer vision and finance. Depending on the data, the performance of wideband separators can sometimes be of the same order, or even better, than that of a neural network or Gaussian mixture model.

	==Español==		==Español==

Sélection de caractéristiques

Pitpitt — Mon, 20 Apr 2026 00:50:00 GMT

← Version précédente		Version du 19 avril 2026 à 20:50
Ligne 18 :		Ligne 18 :
	==Anglais==		==Anglais==
	'''feature selection'''		'''feature selection'''

			Feature selection is a dimensionality reduction method used in machine learning and data processing. It consists, in a high-dimensional space, in finding a subset of relevant variables. In other words, the aim is to minimize the loss of information resulting from the deletion of all other variables.

			See also: feature and characteristic


	==Español==		==Español==

Segmentation

Pitpitt — Mon, 20 Apr 2026 00:48:24 GMT

← Version précédente		Version du 19 avril 2026 à 20:48
Ligne 20 :		Ligne 20 :
	==Anglais==		==Anglais==
	'''tokenization'''		'''tokenization'''

			In automatic natural language processing, the tokenizer segments a text into a list of words, a list of segments or parts of words or symbols, or a sequence of characters. Typically, this is the first step in lexical analysis. See also: token

	==Español==		==Español==

Scientifique des données

Pitpitt — Mon, 20 Apr 2026 00:48:03 GMT

← Version précédente		Version du 19 avril 2026 à 20:48
Ligne 32 :		Ligne 32 :
	==Anglais==		==Anglais==
	''' data scientist '''		''' data scientist '''

			A data scientist is a person who has a sound knowledge of statistics and computer science, and who uses these skills to explore, analyze and extract relevant information from different data sources with the aim of guiding an organization's actions and decision-making.

			The term data scientist therefore refers to a person with in-depth knowledge of data, rather than a scientist whose aim is to extend knowledge in his or her field of study.


	==Español==		==Español==

Science des données

Pitpitt — Mon, 20 Apr 2026 00:46:53 GMT

← Version précédente		Version du 19 avril 2026 à 20:46
Ligne 18 :		Ligne 18 :
	==Anglais==		==Anglais==
	'''data science'''		'''data science'''

			Data science is a multi-disciplinary field at the crossroads of statistics, computer science and application domain.

			Machine learning is generally excluded from data science, as it is considered an autonomous field. This remains an arbitrary choice, as machine learning is essentially data-driven.

			Data science deals with the acquisition, preprocessing, exploration, analysis and data visualization, including the search for relationships and descriptive statistics. Thus defined, data science ends where machine learning proper begins.

			Technically speaking, data science covers all areas related to the exploitation, management, preparation and analysis of massive, structured and unstructured data, with the aim of extracting relevant information or knowledge.


	==Español==		==Español==

Sciences cognitives

Pitpitt — Mon, 20 Apr 2026 00:45:55 GMT

← Version précédente		Version du 19 avril 2026 à 20:45
Ligne 12 :		Ligne 12 :
	==Anglais==		==Anglais==
	'''cognitive sciences'''		'''cognitive sciences'''

			The cognitive sciences are an interdisciplinary group of scientific disciplines such as neuroscience, artificial intelligence, psychology, philosophy, linguistics, anthropology, etc., whose aim is to describe, explain and, where appropriate, model and simulate human, animal or artificial mental mechanisms.

			More generally, the cognitive sciences study and model phenomena as diverse as perception, intelligence, language, memory, attention, reasoning, emotions, the mind and even consciousness.


	==Español==		==Español==

Cognitive sciences

Pitpitt — Mon, 20 Apr 2026 00:45:27 GMT

Page redirigée vers Sciences cognitives

Nouvelle page

#REDIRECTION[[Sciences cognitives]]

Robot conversationnel génératif

Pitpitt — Mon, 20 Apr 2026 00:43:35 GMT

← Version précédente		Version du 19 avril 2026 à 20:43
Ligne 23 :		Ligne 23 :

	'''generative AI chatbot'''		'''generative AI chatbot'''

			A chatbot based on a generative pre-training ranformer. See also: ChatGPT, Gemini and Claude

	==Español==		==Español==

Robot conversationnel

Pitpitt — Mon, 20 Apr 2026 00:42:43 GMT

← Version précédente		Version du 19 avril 2026 à 20:42
Ligne 44 :		Ligne 44 :

	'''conversational agent'''		'''conversational agent'''

			A chatbot is a specialized agent capable of carrying on a conversation. The term conversational agent is also used. It is also defined as a human-machine interface specialized in natural language dialogue with a human user. A chatbot is capable of answering questions, sustaining a simple conversation, or triggering the execution of tasks.

			The conversational agent interprets the user's questions and intentions and provides answers in natural language.

			The conversational agent uses its natural language processing capabilities to answer questions and engage in conversation as a human might.

			Conversational agents are often integrated into automated messaging systems, so as to interact as much as possible like humans. These messaging systems are supported either by a platform of their own, or by existing systems.


	==Español==		==Español==

Robot

Pitpitt — Mon, 20 Apr 2026 00:41:39 GMT

← Version précédente		Version du 19 avril 2026 à 20:41
Ligne 18 :		Ligne 18 :

	'''bot'''		'''bot'''

			The word robot refers to a physical machine, i.e. an automaton, or a programmable software robot (bot) designed to perform a task autonomously.

			A physical robot is a mechatronic device combining mechanics, electronics and computers, designed to perform tasks automatically in a physical environment. It is an advanced automaton capable of perceiving, adapting to its environment and interacting with the objects around it, such as a robot on an automotive assembly line.

			A software robot is an autonomous or semi-autonomous software agent that interacts with its environment. For example, a robot that indexes the Web, or a conversational agent that interacts with a human in natural language.


	==Español==		==Español==

Rétropropagation

Pitpitt — Mon, 20 Apr 2026 00:41:06 GMT

← Version précédente		Version du 19 avril 2026 à 20:41
Ligne 36 :		Ligne 36 :

	'''feedback propagation'''		'''feedback propagation'''

			Backpropagation is a method that enables artificial neural networks to learn. More precisely, error backpropagation is at the heart of the optimization algorithm that enables a neural network to learn from its errors. The general idea is to propagate the error from the output to the input of the neural network in order to adjust the weights or parameters of the network and thus gradually minimize the output error.

	==Español==		==Español==

Réseau récurrent à longue mémoire court terme

Pitpitt — Mon, 20 Apr 2026 00:40:02 GMT

← Version précédente		Version du 19 avril 2026 à 20:40
Ligne 41 :		Ligne 41 :

	'''LSTM'''		'''LSTM'''

			A recurrent network with long short-term memory has several internal mechanisms, such as a memory cell, an input gate, an output gate and a forgetting gate, enabling both short and long dependencies in the data sequences to be taken into account. These gate mechanisms are also designed to alleviate the gradient vanishing problem.

			See also: recurrent network


	==Español==		==Español==

Réseau de neurones récurrent à longue mémoire à court terme

Pitpitt — Mon, 20 Apr 2026 00:39:31 GMT

Destination de redirection modifiée de Réseau de neurones récurrent à longue mémoire court terme en Réseau récurrent à longue mémoire court terme

← Version précédente		Version du 19 avril 2026 à 20:39
Ligne 1 :		Ligne 1 :
	#REDIRECTION [[Réseau ~~de_neurones~~ récurrent à longue mémoire court terme]]		#REDIRECTION [[Réseau récurrent à longue mémoire court terme]]

Réseau récurrent

Pitpitt — Mon, 20 Apr 2026 00:37:49 GMT

← Version précédente		Version du 19 avril 2026 à 20:37
Ligne 22 :		Ligne 22 :

	'''recurrent network'''		'''recurrent network'''

			A recurrent neural network (RNN), or simply recurrent network, is a deep neural network architecture in which there is at least one recurrent connection (loop or cycle) in its structure. In a recurrent network, signals can therefore go backwards via a loop or cycle and feed neurons from a previous layer or from the same layer.

			Recurrent networks can process data of any size. They are particularly suited to the analysis of sequences of data over time, i.e. time series.

			RNRs are used in automatic speech recognition, natural language processing on strings and texts, and machine translation, which are good examples of sequential data processing.


	==Español==		==Español==

Réseau neuronal de graphes

Pitpitt — Mon, 20 Apr 2026 00:35:20 GMT

← Version précédente		Version du 19 avril 2026 à 20:35
Ligne 19 :		Ligne 19 :

	''' GNN '''		''' GNN '''

			A neural network architecture for machine learning from graph-structured data.

	==Español==		==Español==

Réseau neuronal d'espaces d'états structurés

Pitpitt — Mon, 20 Apr 2026 00:34:53 GMT

← Version précédente		Version du 19 avril 2026 à 20:34
Ligne 51 :		Ligne 51 :

	'''state space model learning'''		'''state space model learning'''

			A deep neural network architecture that borrows from recurrent networks, convolutional networks and state space representations to model and process long sequences more efficiently.

	==Español==		==Español==

Réseau neuronal

Pitpitt — Mon, 20 Apr 2026 00:34:18 GMT

← Version précédente		Version du 19 avril 2026 à 20:34
Ligne 16 :		Ligne 16 :
	==Anglais==		==Anglais==
	'''Neural network'''		'''Neural network'''

			An artificial neural network, also known as a neural net or neural network, is a set of interconnected artificial neurons forming a computational architecture capable of solving complex problems by machine learning.

			The main interest of neural networks lies in their architecture and their ability to modify the strength of their connections or parameters by machine learning on data. Neural networks are used to solve complex problems such as pattern recognition or natural language processing.


	==Español==		==Español==

Réseau de neurones à propagation avant

Pitpitt — Mon, 20 Apr 2026 00:33:50 GMT

← Version précédente		Version du 19 avril 2026 à 20:33
Ligne 35 :		Ligne 35 :
	'''forward propagation network'''		'''forward propagation network'''

			A feedforward neural network is an acyclic artificial neural network, as distinct from recurrent neural networks.

			The best-known is the multilayer perceptron, which is an extension of the first artificial neural network. The (simple) perceptron was invented in 1957 by Frank Rosenblatt.

			The forward-propagating neural network was the first and simplest artificial neural network designed. Typically, it had only one hidden layer and was called a Perceptron. In this network, information moves in only one direction, forward, from the input nodes, through the hidden layers (if any) and on to the output nodes. There are no cycles or loops in this network. This is why it is sometimes referred to as a loop-free neural network. When the forward-propagating neural network has several hidden layers, it is usually referred to as a multi-layer perceptron.

	==Español==		==Español==

Réseau de Kolmogorov–Arnold

Pitpitt — Mon, 20 Apr 2026 00:31:53 GMT

← Version précédente		Version du 19 avril 2026 à 20:31
Ligne 24 :		Ligne 24 :

	'''Deep-KAN'''		'''Deep-KAN'''

			A Kolmogorov-Arnold network (KAN) is a deep neural network trained by learning an activation function for each edge or weight of a neural network, rather than a single fixed activation function at the output of the neuron as in a multilayer perceptron. KANs would offer the advantage of better interpretability and higher accuracy.

	==Español==		==Español==

CapsNet

Pitpitt — Mon, 20 Apr 2026 00:31:25 GMT

← Version précédente		Version du 19 avril 2026 à 20:31
Ligne 15 :		Ligne 15 :

	'''CapsNet '''		'''CapsNet '''

			A capsule network is a machine-learning system that is a type of artificial neural network that can be used to better model hierarchical relationships. This approach aims to more closely mimic biological neural organization.

			The idea is to add structures, called capsules, to a convolutional neural network and reuse the outputs of several of these capsules to form more stable (with respect to various perturbations) representations of higher-order capsules. The output is a vector consisting of the probability of an observation and a pose for that observation. This vector is similar to what is done, for example, in classification with localization in convolutional networks.


	==Español==		==Español==

Réseau convolutif

Pitpitt — Mon, 20 Apr 2026 00:30:52 GMT

← Version précédente		Version du 19 avril 2026 à 20:30
Ligne 38 :		Ligne 38 :

	'''CNN'''		'''CNN'''

			A convolutional neural network (CNN) or simply convolutional network is a deep neural network architecture used mainly in computer vision.
			A CNN detects the presence of shapes by applying a local calculation, called convolution, to a group of pixels.

			This convolution calculation is applied to different scales of an image, and enables the progressive identification of image content by association and cross-referencing.

			Convolutional networks have many applications in computer vision, for image and video recognition, but also in sound recognition, recommendation systems and natural language processing.



	==Español==		==Español==

Réseau bayésien

Pitpitt — Mon, 20 Apr 2026 00:29:59 GMT

← Version précédente		Version du 19 avril 2026 à 20:29
Ligne 21 :		Ligne 21 :
	==Anglais==		==Anglais==
	'''Bayesian network'''		'''Bayesian network'''

			In artificial intelligence and statistics, a Bayesian network is a probabilistic model representing a set of random variables in the form of a graph or network.

			A Bayesian network is at once a knowledge representation model, a machine for calculating conditional probabilities and a decision support tool.
			Applications of Bayesian networks include medical and industrial diagnostics, risk analysis, spam detection and exploratory data analysis.


	==Español==		==Español==

Réseau autoattentif

Pitpitt — Mon, 20 Apr 2026 00:28:36 GMT

← Version précédente		Version du 19 avril 2026 à 20:28
Ligne 46 :		Ligne 46 :

	'''self-attentive architecture'''		'''self-attentive architecture'''

			A transformer or self-attention neural network is a high-performance, deep, sequence-to-sequence neural network architecture. It uses the attention mechanism, more precisely self-attention, to replace both recurrence and convolution.

			The self-attention mechanism considerably improves the semantic accuracy of the network, and enables much larger data sets to be processed, as processing can be done more easily in parallel thanks to self-supervised learning.


	==Español==		==Español==

Réseau antagoniste génératif

Pitpitt — Mon, 20 Apr 2026 00:28:01 GMT

← Version précédente		Version du 19 avril 2026 à 20:28
Ligne 24 :		Ligne 24 :

	'''GAN'''		'''GAN'''

			A generative adversarial network or GAN is an unsupervised deep learning algorithm. GAN is a generative model, so it is not designed to classify images, but to generate synthetic images with a high degree of realism.

	==Español==		==Español==

Réseau antagoniste créatif

Pitpitt — Mon, 20 Apr 2026 00:27:38 GMT

← Version précédente		Version du 19 avril 2026 à 20:27
Ligne 24 :		Ligne 24 :

	'''creative GAN'''		'''creative GAN'''

			In deep learning, the creative adversarial network is a variant of the generative adversarial network designed to facilitate artistic creativity.

	==Español==		==Español==

Requête générative

Pitpitt — Mon, 20 Apr 2026 00:26:25 GMT

← Version précédente		Version du 19 avril 2026 à 20:26
Ligne 34 :		Ligne 34 :

	'''prompt'''		'''prompt'''

			In the context of interaction with a chatbot featuring a generative model, a generative query refers to the user input used to interact in natural language.

	==Español==		==Español==

Requête par chaîne de résolution

Pitpitt — Mon, 20 Apr 2026 00:26:02 GMT

← Version précédente		Version du 19 avril 2026 à 20:26
Ligne 41 :		Ligne 41 :

	'''CoT prompt'''		'''CoT prompt'''

			A query that asks a generative chatbot like ChatGPT to respond by breaking down the resolution of a problem into its various stages, for example in the form of a list.

	==Español==		==Español==