Bootstrapping the lexicon: A computational model of infant speech segmentation

Batchelder, Eleanor Olds

doi:10.1016/s0010-0277(02)00002-1

Cited by 57 publications

(71 citation statements)

References 59 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These results seem to conflict with those of several earlier models (Brent, 1999;Venkataraman, 2001;Batchelder, 2002), where systematic undersegmentation was not found even when words were assumed to be independent. However, we argue here that these previous results are misleading.…”

Section: Introductioncontrasting

confidence: 93%

“…We have already shown that the model-based approaches of Venkataraman (2001) and Batchelder (2002) are constrained by their choice of search algorithms; in the following section we demonstrate that the approximate search procedure used by Brent (1999) prevents his learner, too, from identifying the optimal solution under his model. Although in principle one could develop a Bayesian model within the MDL or MBDP frameworks that could account for word-to-word dependencies, the associated search procedures would undoubtedly be even more complex than those required for the current unigram models, and thus even less likely to identify optimal solutions.…”

Section: Systemmentioning

confidence: 89%

“…The kinds of generalizations made by the learner are determined both by the explicit constraints specified by the probabilistic model, and the implicit constraints specified by the search procedure. Examples of this type of learning system are described by Venkataraman (2001) and Batchelder (2002). The models underlying these systems are very similar; we describe only Venkataraman's work in detail.…”

Section: Maximum-likelihood Estimationmentioning

confidence: 99%

“…It should be clear from the previous discussion that this is not the approach advocated here, since it renders constraints implicit and difficult to examine. Batchelder's (2002) maximum-likelihood learning system uses an explicit "external constraint" to penalize lexical items that are too long. This approach is a step in the right direction, but is less mathematically principled than Bayesian modeling, in which a (non-uniform) prior distribution over hypotheses is used within the model itself to constrain learning.…”

Section: Maximum-likelihood Estimationmentioning

confidence: 99%

See 3 more Smart Citations

A Bayesian framework for word segmentation: Exploring the effects of context

2009

View full text Add to dashboard Cite

Since the experiments of Saffran et al. (1996a), there has been a great deal of interest in the question of how statistical regularities in the speech stream might be used by infants to begin to identify individual words. In this work, we use computational modeling to explore the effects of different assumptions the learner might make regarding the nature of words -in particular, how these assumptions affect the kinds of words that are segmented from a corpus of transcribed child-directed speech. We develop several models within a Bayesian ideal observer framework, and use them to examine the consequences of assuming either that words are independent units, or units that help to predict other units. We show through empirical and theoretical results that the assumption of independence causes the learner to undersegment the corpus, with many two-and three-word sequences (e.g. what's that, do you, in the house) misidentified as individual words. In contrast, when the learner assumes that words are predictive, the resulting segmentation is far more accurate. These results indicate that taking context into account is important for a statistical word segmentation strategy to be successful, and raise the possibility that even young infants may be able to exploit more subtle statistical patterns than have usually been considered.

show abstract

Section: Introductioncontrasting

confidence: 93%

Section: Systemmentioning

confidence: 89%

Section: Maximum-likelihood Estimationmentioning

confidence: 99%

Section: Maximum-likelihood Estimationmentioning

confidence: 99%

See 2 more Smart Citations

A Bayesian framework for word segmentation: Exploring the effects of context

2009

View full text Add to dashboard Cite

show abstract

“…Batchelder, 2002). More general and abstract work typically involves computational models capable of learning certain classes of formal languages generated by small artificial grammars (e.g.…”

Section: Introduction a N D O V E R V I E Wmentioning

confidence: 99%

An empirical generative framework for computational modeling of language acquisition

et al. 2010

View full text Add to dashboard Cite

A B S T R A C TThis paper reports progress in developing a computer model of language acquisition in the form of (1) a generative grammar that is (2) algorithmically learnable from realistic corpus data, (3) viable in its large-scale quantitative performance and (4) psychologically real. First, we describe new algorithmic methods for unsupervised learning of generative grammars from raw CHILDES data and give an account of the generative performance of the acquired grammars. Next, we summarize findings from recent longitudinal and experimental work that suggests how certain statistically prominent structural properties of child-directed speech may facilitate language acquisition. We then present a series of new analyses of CHILDES data indicating that the desired properties [*] During the preparation of this paper, Shimon

show abstract

Language Development

MacWhinney¹

2010

The Handbook of Life‐Span Development

View full text Add to dashboard Cite

Bootstrapping the lexicon: A computational model of infant speech segmentation

Cited by 57 publications

References 59 publications

A Bayesian framework for word segmentation: Exploring the effects of context

A Bayesian framework for word segmentation: Exploring the effects of context

An empirical generative framework for computational modeling of language acquisition

Language Development

Contact Info

Product

Resources

About