Recent developments in Transformer language models now allow users to predict the probability of different sentences and to predict missing words more accurately than before. This new information and perspective can be used to form judgments on novel textual emendations and to further quantify existing historical editorial judgments. We examine the importance of analyzing an author’s corpus, and the impact of the Good-Turing theory of frequency estimation when predicting missing words. We will also outline some of the limits of what Transformer language models can do, and how to practically evaluate them.Les développements récents des modèles de langage Transformer permettent désormais à leurs utilisa...
While languages convey significantly different amounts of both information per syllable and syllable...
International audienceSituating his remarks in the theoretical space of "editorial enunciation" and ...
International audienceThe standard metric used to evaluate the performance of Automatic Speech Recog...
People read for various purposes like learning specific skills, acquiring foreign languages, and enj...
Information formulated in natural language is being created at an incredible pace, far more quickly ...
Statistical language modelling may not only be used to uncover the patterns which underlie the compo...
International audienceThis work evaluates quotation finding approaches in the context of ancient Gre...
This paper provides an analysis of all the occurrences of a construction of habeo plus a perfect par...
International audienceThese last years, historians of ancient periods have somehow neglected compute...
The present book can be considered a continuation of two streams in textology initiated by different...
We are so used to speaking in our native language that we take this ability for granted. We think th...
Pie Model for Classical French, for Part-of-Speech and Morphology tags (CATTEX2009-max). Trained on...
In this article we showcase the relevance of corpus evidence in examining potential differences in e...
International audienceMachine translation systems are not reliable enough to be used ''as is'': exce...
Aquesta tesi està dedicada a l'estudi de la utilització de informació morfosintàctica en el marc del...
While languages convey significantly different amounts of both information per syllable and syllable...
International audienceSituating his remarks in the theoretical space of "editorial enunciation" and ...
International audienceThe standard metric used to evaluate the performance of Automatic Speech Recog...
People read for various purposes like learning specific skills, acquiring foreign languages, and enj...
Information formulated in natural language is being created at an incredible pace, far more quickly ...
Statistical language modelling may not only be used to uncover the patterns which underlie the compo...
International audienceThis work evaluates quotation finding approaches in the context of ancient Gre...
This paper provides an analysis of all the occurrences of a construction of habeo plus a perfect par...
International audienceThese last years, historians of ancient periods have somehow neglected compute...
The present book can be considered a continuation of two streams in textology initiated by different...
We are so used to speaking in our native language that we take this ability for granted. We think th...
Pie Model for Classical French, for Part-of-Speech and Morphology tags (CATTEX2009-max). Trained on...
In this article we showcase the relevance of corpus evidence in examining potential differences in e...
International audienceMachine translation systems are not reliable enough to be used ''as is'': exce...
Aquesta tesi està dedicada a l'estudi de la utilització de informació morfosintàctica en el marc del...
While languages convey significantly different amounts of both information per syllable and syllable...
International audienceSituating his remarks in the theoretical space of "editorial enunciation" and ...
International audienceThe standard metric used to evaluate the performance of Automatic Speech Recog...