A Fully Attention-Based Information Retriever

Autor: Correia, Alvaro Henrique Chaim, Silva, Jorge Luiz Moreira, Martins, Thiago de Castro, Cozman, Fabio Gagliardi

Publikováno v: A. H. C. Correia, J. L. M. Silva, T. d. C. Martins and F. G. Cozman, "A Fully Attention-Based Information Retriever," 2018 International Joint Conference on Neural Networks (IJCNN), Rio de Janeiro, Brazil, 2018, pp. 2799-2806

Recurrent neural networks are now the state-of-the-art in natural language processing because they can build rich contextual representations and process texts of arbitrary length. However, recent developments on attention mechanisms have equipped fee

Externí odkaz: http://arxiv.org/abs/1810.09580