scispace - formally typeset
Search or ask a question

“What’s on Your Mind?” - A Literary Dialogue with the Machine-Computer

About: . It has received None citations till now.
References
More filters
Posted Content
TL;DR: Experimental results show the effectiveness of the proposed system design and deployment approach and state-of-the-art machine learning techniques for language model training which enabled high-quality suggestion prediction, and constructed novel serving infrastructure for high-throughput and real-time inference.
Abstract: In this paper, we present Smart Compose, a novel system for generating interactive, real-time suggestions in Gmail that assists users in writing mails by reducing repetitive typing. In the design and deployment of such a large-scale and complicated system, we faced several challenges including model selection, performance evaluation, serving and other practical issues. At the core of Smart Compose is a large-scale neural language model. We leveraged state-of-the-art machine learning techniques for language model training which enabled high-quality suggestion prediction, and constructed novel serving infrastructure for high-throughput and real-time inference. Experimental results show the effectiveness of our proposed system design and deployment approach. This system is currently being served in Gmail.

101 citations

Journal ArticleDOI
24 Feb 2020
TL;DR: In this paper, a typologie des documents nativement numeriques that sont les œuvres litteraires generees par intelligence artificielle, entendant par-la les textes generes par apprentissage machine (machine learning).
Abstract: Dans cet article, je souhaite proposer une typologie des documents nativement numeriques que sont les œuvres litteraires generees par intelligence artificielle, entendant par-la les textes generes par apprentissage machine (machine learning). Je distingue ainsi entre le texte œuvre d’art (dont le caractere litteraire est minore face a un statut d’objet artistique, notamment au travers de la demarche mise en œuvre), le texte edite (dont le caractere litteraire est au contraire mis en avant grâce au travail de reecriture effectue par l’auteur sur le premier jet genere par la machine) et le texte appropriant (dont l’interet a surtout trait au rapport avec les textes utilises pour entrainer les algorithmes, lesquels font les frais d’une appropriation singuliere). Afin d’appuyer cette proposition, je m’appuie notamment sur trois etudes de cas : 1 the Road [Goodwin, 2018], ReRites [Jhave, 2019] et Proust_unlimited [Lebrun, 2018].

11 citations

Book
01 Jan 2009

7 citations