Presentation 2000/11/27
Acquiring word meanings in a small world using neural networks
Mats Nystrand, Kazuhiro Ueda, Naoto Takahashi,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper presents the results from training a neural network on word understanding in short questions about the state of a simulated small world. The neural network receives its input from two sources: 1)a question list and 2)a data structure of the simulated objects and their locations. The output of the network is represented as one or more words from a fixed word list. Back propagation has been used as learning algorithm and the results from several learning tasks are presented. This paper also shows that for a limited range of input a plain non-recurrent three-layer network can learn to answer correctly to questions regarding the position of objects in the small world.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Question & Answering Systems / Neural Networks / Natural Language Processing / Knowledge Acquisition
Paper # TL2000-30
Date of Issue

Conference Information
Committee TL
Conference Date 2000/11/27(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Thought and Language (TL)
Language ENG
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Acquiring word meanings in a small world using neural networks
Sub Title (in English)
Keyword(1) Question & Answering Systems
Keyword(2) Neural Networks
Keyword(3) Natural Language Processing
Keyword(4) Knowledge Acquisition
1st Author's Name Mats Nystrand
1st Author's Affiliation University of Tokyo, Graduate School of International and Interdisciplinary()
2nd Author's Name Kazuhiro Ueda
2nd Author's Affiliation University of Tokyo, Graduate School of International and Interdisciplinary
3rd Author's Name Naoto Takahashi
3rd Author's Affiliation University of Tokyo, Graduate School of International and Interdisciplinary
Date 2000/11/27
Paper # TL2000-30
Volume (vol) vol.100
Number (no) 480
Page pp.pp.-
#Pages 8
Date of Issue