Learning Human-like Knowledge by Singular Value Decomposition: A Progress Report

By T. Landauer, Darrell Laham, Peter W. Foltz. 1997

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Multi-Relational Latent Semantic Analysis (2013, background): “…-th words can be computed by cosine similarity 1 : When used to compare words, one well-known limitation of LSA is that the score captures the general notion of semantic similarity, and is unable to distinguish ﬁne-grained word relations, such as antonyms (Landauer and Laham, 1998; Landauer, 2002).”
Learning Distributed Representations of Concepts Using Linear Relational Embedding (2001, methodology): “Learning Distributed Representations of Concepts using Linear Relational Embedding Alberto Paccanaro Geoffrey Hinton Gatsby Unit”
A Model of Grounded Language Acquisition: Sensorimotor Features Improve Lexical and Grammatical Learning. (2005, methodology): “The work of Landauer and colleagues (e.g., Landauer and Dumais, 1997; Landauer, Laham, & Foltz, 1998 ) provides perhaps the clearest demonstration that word ‘‘meanings’’ can be learned solely from word-to-word relationships (although see Burgess & Lund, 2000; for a different method called HAL).”
GaP: a factor model for discrete data (2004, methodology): “Categories and Subject Descriptors H.3.3 [Information Storage and Retrieval]: Information Search and Retrieval Retrieval Models, Clustering General Terms Algorithms, Experimentation, Measurement Keywords Probabilistic Models, Latent Semantic Analysis, EM Algorithm 1.”
Polarity Inducing Latent Semantic Analysis (2012, background): “For example, latent semantic analysis might assign a high degree of similarity to opposites as well as synonyms ( Landauer and Laham, 1998; Landauer, 2002).,However, as noted by ( Landauer and Laham, 1998; Landauer, 2002), there is no notion of antonymy; words with low or negative cosine scores ar”
Sentence similarity based on semantic nets and corpus statistics (2006, methodology): “The matrix is decomposed by singular value decomposition (SVD) into the product of three other matrices, including the diagonal matrix of singular values [ 19 ].”
Integrating experiential and distributional data to learn semantic representations. (2009, methodology): “In Landauer et al. (1998), it was further reported that LSA performed comparably to children (r ϭ .5) and adults (r ϭ .32) on a word-sorting task.”