Distributional Learning of Appearance

Griffin, Lewis D.; Wahab, M. Husni; Newell, Andrew J.
February 2013
PLoS ONE;Feb2013, Vol. 8 Issue 2, p1
Academic Journal
Opportunities for associationist learning of word meaning, where a word is heard or read contemperaneously with information being available on its meaning, are considered too infrequent to account for the rate of language acquisition in children. It has been suggested that additional learning could occur in a distributional mode, where information is gleaned from the distributional statistics (word co-occurrence etc.) of natural language. Such statistics are relevant to meaning because of the Distributional Principle that ‘words of similar meaning tend to occur in similar contexts’. Computational systems, such as Latent Semantic Analysis, have substantiated the viability of distributional learning of word meaning, by showing that semantic similarities between words can be accurately estimated from analysis of the distributional statistics of a natural language corpus. We consider whether appearance similarities can also be learnt in a distributional mode. As grounds for such a mode we advance the Appearance Hypothesis that ‘words with referents of similar appearance tend to occur in similar contexts’. We assess the viability of such learning by looking at the performance of a computer system that interpolates, on the basis of distributional and appearance similarity, from words that it has been explicitly taught the appearance of, in order to identify and name objects that it has not been taught about. Our experiment tests with a set of 660 simple concrete noun words. Appearance information on words is modelled using sets of images of examples of the word. Distributional similarity is computed from a standard natural language corpus. Our computation results support the viability of distributional learning of appearance.


Related Articles

  • Popularising corpus consultation by language learners and teachers. Chambers, Angela // Language & Computers;2007, Vol. 61 Issue 1, p3 

    Summarising the progress in relation to the use of corpora in language learning made between the first TaLC conference in 1994 and the time of writing their article, McEnery and Wilson (1997: 5) use the term `percolation' to describe how the exploitation of corpora has progressed from the domain...

  • The Potential of Corpus-Informed L2 Pedagogy. Reinhardt, Jonathon // Studies in Hispanic & Lusophone Linguistics;Mar2010, Vol. 3 Issue 1, p239 

    Corpus linguistic methods have led to many revelations about the nature of language use and language learning which are otherwise untenable. To date, however, these findings have not had major impact on L2 (second and foreign language) teaching, except in a few areas like L2 for specialized...

  • Using corpora: from learning to research. Johansson, Stig // Language & Computers;2007, Vol. 61 Issue 1, p17 

    One of the most exciting aspects of corpora is that they may be used to good advantage both in research and teaching. A brief historical sketch is given, stressing the importance of studying language in context. If learners are given access to corpora, they can be encouraged to observe language...

  • CORPORA OF SPOKEN LITHUANIAN. Dabašinskienė, Ineta; Kamandulytė, Laura // Eesti Rakenduslingvistika Ãœhingu Aastaraamat;2009, Vol. 5, p67 

    The paper discusses the development of spoken Lithuanian corpora. In the analytical part longitudinal child language data as well as adult conversations are discussed in view of the issues that occurred during the period of data collection, transcription and coding. The data are transcribed and...

  • Investigating restricted semantic sets in a large general corpus: learning activities for students of English as a foreign language. Coffey, Stephen // Language & Computers;2007, Vol. 61 Issue 1, p161 

    This article describes a type of foreign language learning activity in which concordancing techniques, integrated with more usual classroom activities, are used to raise awareness of certain features of content words in `general English'. In particular, the activity aims to highlight...

  • Using regressions to explore deviations between corpus data and a standard/target: two suggestions. Gries, Stefan Th.; Deshors, Sandra C. // Corpora;2014, Vol. 9 Issue 1, p109 

    The main goal of this study is to develop more appropriate ways to study variation between corpus data that instantiate a linguistic standard or target on the one hand, and corpus data that are compared to that standard, or that represent speakers that may aspire to approximate the target (such...

  • Corpus linguistics and language documentation: challenges for collaboration. Cox, Christopher // Language & Computers;2011, Vol. 73 Issue 1, p239 

    Recent literature in corpus linguistics (e.g., McEnery & Ostler 2000) and language documentation (e.g., Johnson 2004) suggests both disciplines may share natural points of interaction, having in common an interest in the construction and use of permanent collections of diverse linguistic data....

  • Onko lähisukukielen vaikutus suomen ja viron omaksumiseen symmetristä? Korpuspohjaisen tutkimuksen tuloksia ja haasteita. Kaivapalu, Annekatrin; Eslon, Pille // Lähivõrdlusi / Lähivertailuja;2011, Issue 21, p132 

    The paper deals with the influence of closely related first language on the acquisition of Finnish and Estonian. This study is a part of the research project Cross-linguistic influence and second language acquisition: corpus-based research of Tallinn University and is based on parellel use of...

  • NEGOTIATING MEANING IN INTERACTION BETWEEN ENGLISH AND SPANISH SPEAKERS VIA COMMUNICATIVE STRATEGIES. Fernández Dobao, Ana M.; Palacios Martínez, Ignacio M. // Atlantis (0210-6124);jun2007, Vol. 29 Issue 1, p87 

    The main purpose of this paper is to describe and understand how learners and their interlocutors manage to communicate meaning through the use of communication strategies (CSs). Drawing on the collaborative theory of communication, CS episodes are identified and examined to describe how...


Read the Article


Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics