BAM Neural Networks for Machine Translation

スポンサーリンク

概要

• 論文の詳細を見る

• Many problems in machine translation can be characterized as a matter of matching input sequences to output sequences. BAM (bidirectional associative matix) neuralnetworks offer a content addressable encoding of matched pattern pairs in a distributed memory format. In this article, we explore the application of BAM neural networks, to a problem in machine translation. Furthermore, we also apply an extension of the BAM concept, the BAM System, which overcomes the memory limitations of the original BAM model

• 一般社団法人情報処理学会の論文
• 1993-03-01

著者

• Seki Hirohisa

  Nagoya Institute Of Technology

• Hidenori Itoh

  Nagoya Institute Of Technology

• ToddWarren Law

  Nagoya Institute of Technology

関連論文

• Constrained Knot Representation and Its Characteristics : 紐デザイン処理(3)
• BAM Neural Networks for Machine Translation

スポンサーリンク