TERMIUM Plus^®

Record 1,  Textual support,  English

Record number: 1, Textual support number: 1 CONT

Self's vision of formalizing teaching in terms of a student model is given a sharp focus in his work on concept teaching(Self, 1977). It differs from other projects by being a formal experiment on teaching rather than a teaching system. The domain is the acquisition of simple conjunctive concepts in a relational language close to first-order logic. An example of the concepts taught by Self's system would be R(x) T(y)("There is a rectangle and no triangle"). The target concept is taught through successive variable-free examples classified as positive or negative instances. For example, the tutor could say that R(a) T(b)("There is a rectangle a and a triangle b") is a negative example. Since the order in which the sequence of instances is presented greatly influences the time the student takes to converge on the correct concept, the idea is for the teacher to use a learning model to optimize its selection of examples.  1, record 1, English, - concept%20teaching

Record 1,  Textual support,  French

Record number: 1, Textual support number: 1 OBS

Voir étude de concepts, évaluation des concepts, test de concepts.  3, record 1, French, - enseignement%20de%20type%20conceptuel

Record number: 1, Textual support number: 2 OBS

Comparer à apprentissage de concepts : apprentissage visant à trouver une fonction de reconnaissance qui permet de distinguer les éléments appartenant au concept de ceux n'en faisant pas partie.  3, record 1, French, - enseignement%20de%20type%20conceptuel

Record 1,  Textual support,  Spanish

Record 2,  Textual support,  English

Record number: 2, Textual support number: 1 OBS

When parameter transfer is used, some parameters from the model trained on the source domain are reused in the model intended for the target domain.  2, record 2, English, - parameter%20transfer

Record 2,  Textual support,  French

Record number: 2, Textual support number: 1 OBS

Lorsqu'on fait appel au transfert de paramètres, certains paramètres du modèle entraîné à partir du domaine source sont réutilisés dans le modèle destiné à être utilisé pour le domaine cible.  2, record 2, French, - transfert%20de%20param%C3%A8tres

Record 2,  Textual support,  Spanish

Record 3,  Textual support,  English

Record number: 3, Textual support number: 1 OBS

In transfer learning, the knowledge a source model has acquired by training on data from a source domain to perform a source task is transferred to a target model. This knowledge is used by the target model as a starting point to train on data from the target domain and perform the target task.  2, record 3, English, - target%20domain

Record 3,  Textual support,  French

Record number: 3, Textual support number: 1 OBS

En apprentissage par transfert, les connaissances qu'un modèle source a acquises en s'entraînant au moyen de données d'un domaine source en vue de réaliser une tâche source sont transférées à un modèle cible. Le modèle cible utilise ces connaissances comme point de départ pour s'entraîner au moyen de données du domaine cible et effectuer la tâche cible.  2, record 3, French, - domaine%20cible

Record 3,  Textual support,  Spanish

Record 4,  Textual support,  English

Record number: 4, Textual support number: 1 OBS

In transfer learning, the knowledge a source model has acquired by training on data from a source domain to perform a source task is transferred to a target model. This knowledge is used by the target model as a starting point to train on data from the target domain and perform the target task.  2, record 4, English, - target%20model

Record 4,  Textual support,  French

Record number: 4, Textual support number: 1 OBS

En apprentissage par transfert, les connaissances qu'un modèle source a acquises en s'entraînant au moyen de données d'un domaine source en vue de réaliser une tâche source sont transférées à un modèle cible. Le modèle cible utilise ces connaissances comme point de départ pour s'entraîner au moyen de données du domaine cible et effectuer la tâche cible.  2, record 4, French, - mod%C3%A8le%20cible

Record 4,  Textual support,  Spanish

Record 5,  Textual support,  English

Record number: 5, Textual support number: 1 OBS

Inductive transfer learning is used when the target task is different from, yet related to, the source task. This technique can be applied whether the source and target domains are the same or not, provided that the target domain contains some labelled data. As for the source domain, it may contain labelled data or not.  2, record 5, English, - inductive%20transfer%20learning

Record 5,  Textual support,  French

Record number: 5, Textual support number: 1 OBS

L'apprentissage par transfert inductif est utilisé lorsque la tâche cible est différente de la tâche source, tout en lui étant apparentée. Cette technique peut être employée que les domaines source et cible soient identiques ou non, à condition que le domaine cible contienne une certaine quantité de données étiquetées. Le domaine source peut quant à lui comporter ou non des données étiquetées.  2, record 5, French, - apprentissage%20par%20transfert%20inductif

Record 5,  Textual support,  Spanish

Record 6,  Textual support,  English

Record number: 6, Textual support number: 1 OBS

In transfer learning, the knowledge a source model has acquired by training on data from a source domain to perform a source task is transferred to a target model. This knowledge is used by the target model as a starting point to train on data from the target domain and perform the target task.  2, record 6, English, - source%20domain

Record 6,  Textual support,  French

Record number: 6, Textual support number: 1 OBS

En apprentissage par transfert, les connaissances qu'un modèle source a acquises en s'entraînant au moyen de données d'un domaine source en vue de réaliser une tâche source sont transférées à un modèle cible. Le modèle cible utilise ces connaissances comme point de départ pour s'entraîner au moyen de données du domaine cible et effectuer la tâche cible.  2, record 6, French, - domaine%20source

Record 6,  Textual support,  Spanish

Record 7,  Textual support,  English

Record number: 7, Textual support number: 1 OBS

In transfer learning, the knowledge a source model has acquired by training on data from a source domain to perform a source task is transferred to a target model. This knowledge is used by the target model as a starting point to train on data from the target domain and perform the target task.  2, record 7, English, - source%20task

Record 7,  Textual support,  French

Record number: 7, Textual support number: 1 OBS

En apprentissage par transfert, les connaissances qu'un modèle source a acquises en s'entraînant au moyen de données d'un domaine source en vue de réaliser une tâche source sont transférées à un modèle cible. Le modèle cible utilise ces connaissances comme point de départ pour s'entraîner au moyen de données du domaine cible et effectuer la tâche cible.  2, record 7, French, - t%C3%A2che%20source

Record 7,  Textual support,  Spanish

Record 8,  Textual support,  English

Record number: 8, Textual support number: 1 CONT

Domain adaptation is a subfield of machine learning that focuses on adapting a model trained on one domain(source domain) to perform well on a different, but related domain(target domain). The main challenge in domain adaptation is to overcome the domain shift, which is the difference in data distributions between the source and target domains.  2, record 8, English, - domain%20adaptation

Record 8,  Textual support,  French

Record number: 8, Textual support number: 1 CONT

L'adaptation de domaine aborde le défi des distributions de données différentes entre les domaines source et cible [en] visant à rendre le modèle plus adaptable au domaine cible [...]  2, record 8, French, - adaptation%20de%20domaine

Record 8,  Textual support,  Spanish

Record 9,  Textual support,  English

Record number: 9, Textual support number: 1 OBS

Transductive transfer learning is used when the source and target tasks are of the same nature, but the source and target domains are different, yet related. This technique applies when the target domain does not include labelled data, while the source domain does. The source and target domains can differ in terms of their feature space or data distribution.  2, record 9, English, - transductive%20transfer%20learning

Record 9,  Textual support,  French

Record number: 9, Textual support number: 1 OBS

L'apprentissage par transfert transductif est utilisé lorsque les tâches source et cible sont de même nature, mais que les domaines source et cible sont différents, tout en étant apparentés. Cette technique est employée dans les cas où le domaine cible ne comporte pas de données étiquetées, alors que le domaine source en comporte. Les domaines source et cible peuvent être différents en ce qui a trait à leur espace de caractéristiques ou à la distribution de leurs données.  2, record 9, French, - apprentissage%20par%20transfert%20transductif

Record 9,  Textual support,  Spanish

Record 10,  Textual support,  English

Record number: 10, Textual support number: 1 OBS

In transfer learning, the knowledge a source model has acquired by training on data from a source domain to perform a source task is transferred to a target model. This knowledge is used by the target model as a starting point to train on data from the target domain and perform the target task.  2, record 10, English, - source%20model

Record 10,  Textual support,  French

Record number: 10, Textual support number: 1 OBS

En apprentissage par transfert, les connaissances qu'un modèle source a acquises en s'entraînant au moyen de données d'un domaine source en vue de réaliser une tâche source sont transférées à un modèle cible. Le modèle cible utilise ces connaissances comme point de départ pour s'entraîner au moyen de données du domaine cible et effectuer la tâche cible.  2, record 10, French, - mod%C3%A8le%20source

Record 10,  Textual support,  Spanish

Record 11,  Textual support,  English

Record number: 11, Textual support number: 1 OBS

In transfer learning, the knowledge a source model has acquired by training on data from a source domain to perform a source task is transferred to a target model. This knowledge is used by the target model as a starting point to train on data from the target domain and perform the target task.  2, record 11, English, - target%20task

Record 11,  Textual support,  French

Record number: 11, Textual support number: 1 OBS

En apprentissage par transfert, les connaissances qu'un modèle source a acquises en s'entraînant au moyen de données d'un domaine source en vue de réaliser une tâche source sont transférées à un modèle cible. Le modèle cible utilise ces connaissances comme point de départ pour s'entraîner au moyen de données du domaine cible et effectuer la tâche cible.  2, record 11, French, - t%C3%A2che%20cible

Record 11,  Textual support,  Spanish

Record 12,  Textual support,  English

Record number: 12, Textual support number: 1 DEF

Didactic level to be reached by tutoring, or concept teaching.  1, record 12, English, - target%20level

Record number: 12, Textual support number: 1 CONT

From the target level perspective, hints or pieces of advice are fundamentally different from explanations, and correction are likewise different from remediation, in that they only address behavior. Similarly, simple demonstration and traces of reasoning are restricted to exposing behavior in the domain without providing other forms of support for learning. Finally, simulations only expose the behavior of objects. Such didactic operations all require a good deal of interpretation by the student in order to be converted into useful knowledge.  1, record 12, English, - target%20level

Record 12,  Textual support,  French

Record number: 12, Textual support number: 1 DEF

Niveau de connaissances à atteindre par l'apprenant suite à une session d'apprentissage.  1, record 12, French, - niveau%20cible

Record 12,  Textual support,  Spanish

Record 13,  Textual support,  English

Record number: 13, Textual support number: 1 OBS

The Joint Unmanned Surveillance and Target Acquisition System(JUSTAS) will complement existing reconnaissance, surveillance, and target acquisition capabilities, increase maritime and arctic domain awareness and provide precision force application in support of deployed Land and Special Operations Forces.  5, record 13, English, - Joint%20Unmanned%20Surveillance%20and%20Target%20Acquisition%20System

Record 13,  Textual support,  French

Record number: 13, Textual support number: 1 OBS

Le Système interarmées de surveillance et d’acquisition d’objectifs au moyen de véhicules aériens sans pilote (JUSTAS) sera complémentaire aux capacités actuelles de reconnaissance, de surveillance et d’acquisition d’objectifs, accroîtra la connaissance de la situation maritime, notamment dans l’Arctique, et favorisera la précision dans l’emploi de la force en appui à la Force terrestre ainsi qu’aux Forces d’opérations spéciales.  3, record 13, French, - Syst%C3%A8me%20interarm%C3%A9es%20de%20surveillance%20et%20d%26rsquo%3Bacquisition%20d%26rsquo%3Bobjectifs%20au%20moyen%20de%20v%C3%A9hicules%20a%C3%A9riens%20sans%20pilote

Record 13,  Textual support,  Spanish

Record 14,  Textual support,  English

Record number: 14, Textual support number: 1 CONT

The interaction of the HIV-1 Rev protein with its target, the Rev responsive element(RRE). The Rev regulatory protein of HIV Rev is modular protein with a N-terminal RNA binding motif and near C-terminal activation domain.  1, record 14, English, - Rev%20Responsive%20Element

Record 14,  Textual support,  French

Record number: 14, Textual support number: 1 CONT

Chaque étape du cycle viral peut constituer une cible d'intervention antivirale. Les agent antiviraux en évaluation dans le monde sont: CD4 soluble, anticorps anti-Env intracellulaires ou secrétés, ribozymes, anticorps intracellulaires antiscriptase inverse (anti-RTase) IFN, leurres de séquence TAR, cible de TAT, inhibiteurs transdominants de TAT, leurres de la séquence RRE de réponse à REV, ARN antisens, inhibiteurs transdominants GAG, CD4 intracellulaires, anticorps anti-ENV "suicide-induit".  1, record 14, French, - s%C3%A9quence%20REE

Record 14,  Textual support,  Spanish

Record 15,  Textual support,  English

Record number: 15, Textual support number: 1 OBS

... the conceptual system must have concepts or rules(schema and transformations) which permit a mapping from the representational domain(lines, patches of color on paper) to the target domain(recognizable events, 3-D objects, etc.).  1, record 15, English, - representational%20domain

Record 15,  Textual support,  French

Record number: 15, Textual support number: 1 OBS

Équivalents proposés par un psycholinguiste, Université du Québec à Trois-Rivières.  1, record 15, French, - domaine%20de%20la%20repr%C3%A9sentation

Record 15,  Textual support,  Spanish

Record 16,  Textual support,  English

Record number: 16, Textual support number: 1 OBS

... the conceptual system must have concepts or rules(schema and transformations) which permit a mapping from the representational domain(lines, patches of color on paper) to the target domain(recognizable events, 3-D objects, etc.).  1, record 16, English, - target%20domain

Record 16,  Textual support,  French

Record number: 16, Textual support number: 1 OBS

Équivalent proposé par un psycholinguiste, Université du Québec à Trois-Rivières.  1, record 16, French, - domaine%20cible

Record 16,  Textual support,  Spanish

Record 17,  Textual support,  English

Record number: 17, Textual support number: 1 CONT

Relations, or predicates connecting several objects or concepts, are mapped identically from one domain to the other under a prespecified object-object correspondence. After identical first-order relations have been mapped to related corresponding objects in a target situation, second-order predicates, such as causal links between relations, are also mapped.  2, record 17, English, - map

Record 17,  Textual support,  French

Record number: 17, Textual support number: 1 CONT

La détermination des caractéristiques à mettre en correspondance peut se faire selon différentes heuristiques visant à contraindre l'espace des correspondances possibles; l'objectif est de privilégier les caractéristiques qui ont le plus d'utilité et de chance d'être préservées ou transposables d'une analogie à l'autre.  2, record 17, French, - mettre%20en%20correspondance

Record 17,  Textual support,  Spanish

Record 18,  Textual support,  English

Record number: 18, Textual support number: 1 CONT

A very original set of experiments carried out by the Unit of Molecular Biology of Gene Expression concerns another protein essential for mouse development that is also a transcription factor. This factor known as KBF2 or RBP-Jk is homologous to a Drosophila gene product, Su(H)(Suppressor of Hairless), which belongs to the Notch signal transduction cascade involved in regulating the competence of a cell to respond to specific developmental cues. A likely function of the Su(H) protein is to serve as an anchor on DNA for non DNA-binding transcriptional activators. This function is conserved in the Notch pathway : the activation of this transmembrane receptor likely involves a proteolytic cleavage step which releases its intracellular domain; following nuclear translocation, it activates its target genes by binding to the Su(H) protein, which is already bound to several sites in the target promoters.  2, record 18, English, - Su%28H%29%20protein

Record 18,  Textual support,  French

Record number: 18, Textual support number: 1 CONT

La protéine Su(H), impliquée dans l'activité de transmission de ce signal, semble capable de s'associer à la région intracellulaire de Notch. C'est un facteur de transcription qui se lie à certaines séquences localisées dans les promoteurs de plusieurs des gènes cibles de Notch.  1, record 18, French, - prot%C3%A9ine%20Su%28H%29

Record 18,  Textual support,  Spanish

Record 19,  Textual support,  English

Record number: 19, Textual support number: 1 DEF

The identification of the relevant parts of the source to map to the appropriate parts of the target.  2, record 19, English, - analogical%20mapping

Record number: 19, Textual support number: 1 CONT

We have replaced bottom-up matching with an approach based on analogical mapping. Such an approach uses a set of heuristics to delimit what is to be "imported" from a base to a target domain at a given time. One such heuristic involves mapping previously formed abstractions, such as those embodying causal and planning rules.  3, record 19, English, - analogical%20mapping

Record 19,  Textual support,  French

Record number: 19, Textual support number: 1 CONT

Apprentissage par analogie. L'objectif est de privilégier les caractéristiques qui ont le plus d'utilité et de chance d'être préservées ou transposables d'une analogie à l'autre. Les travaux de Gentner ont mis en évidence que les mises en correspondance de systèmes de relations sont nettement plus puissantes à cet égard.  1, record 19, French, - mise%20en%20correspondance%20analogique

Record 19,  Textual support,  Spanish

Record 20,  Textual support,  English

Record number: 20, Textual support number: 1 CONT

... stated correspondences must be presented along with a target domain situation described in terms of a plausible situation in the familiar domain.  1, record 20, English, - target%20situation

Record 20,  Textual support,  French

Record number: 20, Textual support number: 1 CONT

La situation inconnue pour laquelle on infère des propriétés est appelée analogue cible; la situation similaire connue est appelée analogue source.  1, record 20, French, - analogue%20cible

Record 20,  Textual support,  Spanish

Record 21,  Textual support,  English

Record number: 21, Textual support number: 1 CONT

Analogies generally include statements suggesting correspondences between classes of objects in the domains to be related. To be useful, these stated correspondences must be presented along with a target domain situation described in terms of a plausible situation in the familiar domain.  1, record 21, English, - correspondence

Record 21,  Textual support,  French

Record number: 21, Textual support number: 1 CONT

Il faut ensuite élaborer une correspondance entre l'analogue source reconnu et l'analogue cible, de façon à déterminer les caractéristiques utiles à transférer de l'un à l'autre (...).  1, record 21, French, - correspondance

Record 21,  Textual support,  Spanish

Record 22,  Textual support,  English

Record number: 22, Textual support number: 1 CONT

This approach... as a model for analogical learning... presupposes that well-defined, bounded representational models of the situations in both the base(i. e., familiar) domain and the target domain are available as inputs.  1, record 22, English, - analogical%20base%20domain

Record 22,  Textual support,  French

Record number: 22, Textual support number: 1 CONT

Apprentissage par analogie. (...) Le correspondant dans le domaine cible d'un prédicat de comparaison dans le domaine source doit pouvoir porter sur des objets comparables dans le domaine cible.  1, record 22, French, - domaine%20source

Record 22,  Textual support,  Spanish

Record 23,  Textual support,  English

Record number: 23, Textual support number: 1 CONT

The stratagem used in GUIDON is to compile tutorial strategies into didactic packages that are domain-independent, but are specialized to a specific representational scheme for the target expertise. Didactic strategies can thus use the representational syntax as an anchor into the domain. We have noted the methodological importance of this approach in motivating the formation of abstractions.  1, record 23, English, - didactic%20package

Record number: 23, Textual support number: 1 OBS

See expert system shell.  2, record 23, English, - didactic%20package

Record 23,  Textual support,  French

Record number: 23, Textual support number: 1 DEF

Système essentiel d'un didacticiel intelligent, progiciel relativement indépendant du domaine à enseigner.  1, record 23, French, - g%C3%A9n%C3%A9rique%20didactique

Record 23,  Textual support,  Spanish

Record 24,  Textual support,  English

Record number: 24, Textual support number: 1 DEF

Problem which has the five following components : a semantic domain, a syntactic domain, a mapping, a target object, and an oracle.  1, record 24, English, - inductive%20inference%20problem

Record 24,  Textual support,  French

Record 24,  Textual support,  Spanish