TERMIUM Plus^®

Record 1,  Textual support,  English

Record number: 1, Textual support number: 1 CONT

The objective of the changes made in the gain parameter is to minimize the value of the functional given by(the) equation. However, the relationship between the gain and the functional is generally not known a priori. Consequently, we must rely on search techniques to find the value of gain that minimizes the functional, J. To do this, a small change is made in the gain and the error change is observed. If the value of the functional decreases, we continue to change the gain in the same direction. This process(is) a realization of the gradient search or steepest descent technique....  2, record 1, English, - gradient%20search

Record 1,  Textual support,  French

Record number: 1, Textual support number: 1 OBS

Descente de plus grande pente : recherche itérative dont l'algorithme développe le meilleur sommet pendant de l'état antérieur jusqu'à la rencontre d'une feuille ou d'un père de feuilles.  2, record 1, French, - recherche%20du%20gradient

Record 1,  Textual support,  Spanish

Record 2,  Textual support,  English

Record number: 2, Textual support number: 1 CONT

Stochastic gradient algorithms are widely used for large-scale learning and inference problems. However, their use in practice is typically guided by heuristics and trial-and-error rather than rigorous, generalizable theory.  5, record 2, English, - stochastic%20gradient%20algorithm

Record number: 2, Textual support number: 2 CONT

Stochastic gradient descent(SGD) and its variants have established themselves as the go-to algorithms for large-scale machine learning problems with independent samples due to their generalization performance and intrinsic computational advantage.  3, record 2, English, - stochastic%20gradient%20algorithm

Record 2,  Textual support,  French

Record 2,  Textual support,  Spanish

Record 3,  Textual support,  English

Record number: 3, Textual support number: 1 CONT

The dynamics of SGD [stochastic gradient descent](online) learning by two-layer neural networks in the teacher-student setup was studied in a series of classic papers … from the statistical physics community …  1, record 3, English, - two%2Dlayer%20neural%20network

Record 3,  Textual support,  French

Record number: 3, Textual support number: 1 CONT

word2vec [...] est un réseau neuronal à deux couches qui est entraîné d'une manière supervisée dans le but de produire une représentation distributionnelle des mots.  3, record 3, French, - r%C3%A9seau%20de%20neurones%20%C3%A0%20deux%20couches

Record 3,  Textual support,  Spanish

Record 4,  Textual support,  English

Record number: 4, Textual support number: 1 CONT

The optimum descent gradient in all segments of helicopter approach procedures is 400 feet per nautical mile.  1, record 4, English, - optimum%20descent%20gradient

Record 4,  Textual support,  French

Record number: 4, Textual support number: 1 CONT

La pente de descente optimale dans tous les segments d'une procédure d'approche pour hélicoptère est de 400 pieds par mille nautique.  1, record 4, French, - pente%20de%20descente%20optimale

Record 4,  Textual support,  Spanish

Record 5,  Textual support,  English

Record number: 5, Textual support number: 1 DEF

The situation where the system always moves downward in the energy function.  2, record 5, English, - gradient%20descent

Record number: 5, Textual support number: 1 CONT

One standard optimization technique is to use gradient descent : The values of the variables in the problem are modified in whatever direction reduces the cost function(energy). For hard problems, gradient descent gets stuck at local minima that are not globally optimal.  3, record 5, English, - gradient%20descent

Record 5,  Textual support,  French

Record number: 5, Textual support number: 1 CONT

La surface de la fonction d'erreur présente des caractéristiques peu satisfaisantes pour effectuer une descente de gradient : minima locaux, qui empêchent la convergence de l'algorithme, plateaux où les pentes sont très faibles, etc.  1, record 5, French, - descente%20de%20gradient

Record 5,  Textual support,  Spanish

Record 6,  Textual support,  English

Record number: 6, Textual support number: 1 OBS

The net flight path is that attainable from the expected gradient of climb or descent diminished by 0. 2 per cent [p. 50-3. 3, note]  1, record 6, English, - net%20flight%20path

Record 6,  Textual support,  French

Record number: 6, Textual support number: 1 OBS

La trajectoire nette de vol est la trajectoire qui peut être réalisée à partir de la pente de montée ou de descente escomptée diminuée de 0,2% [p. 50-3.3, note]  1, record 6, French, - trajectoire%20nette%20de%20vol

Record 6,  Textual support,  Spanish