Sparse Distributed Representations for Words with Thresholded Independent Component Analysis

— We show that independent component analysis (ICA) can be used to find distributed representations for words that can be further processed by thresholding to produce sparse representations. The applicability of the thresholded ICA representation is compared to singular value decomposition (SVD) in a multiple choice vocabulary task with three data sets.

• Type: Conference
• Year: 2007
• Where: IJCNN
• Authors: Jaakko J. Väyrynen, Lasse Lindqvist, Timo Honkela
• Keywords: IJCNN 2007, Independent Component Analysis, Singular Value Decomposition, Thresholded Ica Representation