What is the Dimensionality of Human Semantic Space?
Perspectives in neural computingPublished 1 January 2001
Will Lowe
Citations20
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
This paper provides a replication of McDonald and Lowe’s results in two dimensions using the Generative Topographic Mapping [2], a statistically motivated neural network architecture for topographic maps.
Abstract
McDonald and Lowe [15] showed that cosines in a semantic space of several hundred dimensions reflect human priming results for a wide range of semantic and associatively related words [16, Exp.2]. Previously, Lowe [II, 10] argued that the intrinsic dimensionality of semantic space is much lower, and that high-dimensional structure can be effectively captured in just two dimensions as the surface of a neural map. This paper provides a replication of McDonald and Lowe's results in two dimensions using the Generative Topographic Mapping [2], a statistically motivated neural network architecture for topographic maps.
Keywords
Computer Science
Springer series in information sciencesSelf-Organizing Maps
10,377 Citations1997Teuvo Kohonen
The mathematical preliminaries, background, basic ideas, and implications of the Self-Organising Map algorithm are expounded in a manner which is accessible without prior expert knowledge.
Psychological ReviewA solution to Plato's problem: The latent semantic analysis theory of acquisition, induction, and representation of knowledge.
6,039 Citations1997Thomas K. Landauer, Susan Dumais
A new general theory of acquired similarity and knowledge representation, latent semantic analysis (LSA), is presented and used to successfully simulate such learning and several other psycholinguistic phenomena.
Neural ComputationGTM: The Generative Topographic Mapping
1,389 Citations1998Chris Bishop, Markus Svensén +1 more
A form of nonlinear latent variable model called the generative topographic mapping, for which the parameters of the model can be determined using the expectation-maximization algorithm, is introduced.
Biological CyberneticsSelf-organizing semantic maps
623 Citations1989Helge Ritter, Teuvo Kohonen
Self-organized formation of topographic maps for abstract data, such as words, is demonstrated and it is argued that a similar process may be at work in the brain.
Dimensionality reduction by random mapping: fast similarity computation for clustering
386 Citations2002Samuel Kaski
It is demonstrated that the document classification accuracy obtained after the dimensionality has been reduced using a random mapping method will be almost as good as the original accuracy if the final dimensionality is sufficiently large.
Discourse ProcessesExplorations in context space: Words, sentences, discourse
280 Citations1998Curt Burgess, Kay Livesay +1 more
A computational model of high‐dimensional context space, the Hyperspace Analog to Language (HAL), is described and simulation evidence that HAL's vector representations can provide sufficient information to make semantic, grammatical, and abstract distinctions is presented.
Grammatical Trigrams: A Probabilistic Model of Link Grammar
136 Citations1992John Lafferty, Saniel Sleator +1 more
An algorithm for determining maximum-likelihood estimates of the parameters of these language models, which include the familiar $n$-gram models as a natural subclass, are described.
eScholarship (California Digital Library)Semantic and Associative Priming in High-Dimensional Semantic Space
77 Citations1995Lund, Kevin, Burgess, Curt +1 more
Edinburgh Research Archive (University of Edinburgh)The Direct Route: Mediated Priming in Semantic Space
62 Citations2000Will Lowe, Scott D. McDonald
Rodopi eBooksEnglish language corpora : design, analysis and exploitation : papers from the thirteenth International Conference on English Language Research on Computerized Corpora, Nijmegen 1992
47 Citations1993Jan Aarts, Pieter de Haan +1 more
Neural ComputationElastic Net Model of Ocular Dominance: Overall Stripe Pattern and Monocular Deprivation
43 Citations1994Geoffrey J. Goodhill, David Willshaw
It is shown for this model that the overall pattern of stripes produced is strongly influenced by the shape of the cortex, and stripes with a global order similar to that seen biologically can be produced under appropriate conditions.
Modelling Functional Priming and the Associative Boost
32 Citations2022Scott A. McDonald, Will Lowe
Why previous corpus-based approaches to lexical semantics have failed to produce similar results is investigated, and it is argued that vector representations partly encode temporal co-occurrence information.
Meaning and the mental lexicon
20 Citations1997Will Lowe
A network model of the mental lexicon and its formation is presented, similar to semantic networks first described by [Collins and Loftus, 1975], but is created automatically.
Perspectives in neural computingSemantic Representation and Priming in a Self-organizing Lexicon
15 Citations1998Will Lowe
This paper presents a model of the mental lexicon and its formation, based on the self-organizing neural network, and suggests a novel explanation of semantic priming based on topographic organization and co-occurrence statistics.
Edinburgh Research Archive (University of Edinburgh)Topographic maps of semantic space
10 Citations2001William Lowe
This work develops and tests the hypothesis that similarity of meaning derives from substitution regularities in the linguistic environment, represented topographically in the brain, and develops a general mathematical theory of semantic space models and uses this to motivate a set of new methods for high-dimensional space construction.