Weak semantic context helps phonetic learning in a model of infant language acquisition

Learning phonetic categories is one of the first steps to learning a language, yet is hard to do using only distributional phonetic information. Semantics could potentially be useful, since words with different meanings have distinct phonetics, but it is unclear how many word meanings are known to infants learning phonetic categories. We show that attending to a weaker source of semantics, in the form of a distribution over topics in the current context, can lead to improvements in phonetic category learning. In our model, an extension of a previous model of joint word-form and phonetic category inference, the probability of word-forms is topic-dependent, enabling the model to find significantly better phonetic vowel categories and word-forms than a model with no semantic knowledge.

• Publication year

  2014

• Venue

  Annual Meeting of the Association for Computational Linguistics

• Publication date

  2014-06-01

• Fields of study

  Linguistics, Computer Science

• Identifiers
  DOI 10.3115/v1/P14-1101
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• A Rudimentary Lexicon and Semantics Help Bootstrap Phoneme Acquisition

  2014cited by this paper
• Word-level information influences phonetic learning in adults and infants.

  2013cited by this paper
• A role for the developing lexicon in phonetic category acquisition.

  2013cited by this paper
• Propose but verify: fast mapping meets cross-situational word learning.

  2013cited by this paper
• A Joint Learning Model of Word Segmentation, Lexical Acquisition, and Phonetic Variability

  2013cited by this paper
• A Single-Stage Approach to Learning Phonological Categories: Insights From Inuktitut

  2013cited by this paper
• Developmental Science, 6

  2012cited by this paper
• of the Annual Meeting of the Cognitive Science Society Title Relating Activity Contexts to Early Word Learning in Dense Longitudinal Data

  2012cited by this paper
• Six-Month-Olds Comprehend Words That Refer to Parts of the Body.

  2012cited by this paper
• At 6–9 months, human infants know the meanings of many common nouns

  2012cited by this paper
• Distributional Learning of Vowel Categories Is Supported by Prosody in Infant-Directed Speech

  2012cited by this paper
• How words can and cannot be learned by observation

  2011cited by this paper
• Word Learning in 6-Month-Olds: Fast Encoding–Weak Retention

  2011cited by this paper
• Prosody guides the rapid mapping of auditory word forms onto visual objects in 6-mo-old infants

  2011cited by this paper
• Twelve-Month-Olds Know Their Cups From Their Keps and Tups.

  2010cited by this paper
• Two Decades of Unsupervised POS Induction: How Far Have We Come?

  2010cited by this paper
• Grundzüge der Phonologie

  2010cited by this paper
• Learning words' sounds before learning how words sound: 9-month-olds use distinct objects as cues to categorize speech information.

  2009cited by this paper
• Statistical learning of phonetic categories: insights from a computational approach.

  2009cited by this paper
• Contributions of infant word learning to language development

  2009cited by this paper
• Visual speech contributes to phonetic learning in 6-month-old infants.

  2008cited by this paper
• Statistical phonetic learning in infants: facilitation and feature generalization.

  2008cited by this paper
• Infants rapidly learn word-referent mappings via cross-situational statistics.

  2008cited by this paper
• Unsupervised learning of vowel categories from infant-directed speech

  2007cited by this paper
• Rapid Word Learning Under Uncertainty via Cross-Situational Statistics

  2007cited by this paper
• The effect of distributional information on children's use of phonemic contrasts

  2007cited by this paper
• V-Measure: A Conditional Entropy-Based External Cluster Evaluation Measure

  2007cited by this paper
• A Hierarchical Bayesian Language Model Based On Pitman-Yor Processes

  2006cited by this paper
• Word Learning

  2005cited by this paper
• The basis of preference for lexical words in 6-month-old infants

  2003cited by this paper
• Hierarchical Topic Models and the Nested Chinese Restaurant Process

  2003cited by this paper
• Investigating the role of infant-directed speech with a computer model

  2003cited by this paper
• Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics (ACL)

  2002cited by this paper
• Infant sensitivity to distributional information can affect phonetic discrimination.

  2002cited by this paper
• Intersensory Redundancy and 7-Month-Old Infants' Memory for Arbitrary Syllable-Object Relations

  2001cited by this paper
• The role of exposure to isolated words in early vocabulary development.

  2001cited by this paper
• Markov Chain Sampling Methods for Dirichlet Process Mixture Models

  2000cited by this paper
• The Infinite Gaussian Mixture Model

  1999influential reference
• Some Beginnings of Word Comprehension in 6-Month-Olds

  1999cited by this paper
• Infants listen for more phonetic detail in speech perception than in word-learning tasks

  1997cited by this paper
• Infants' memory for spoken words.

  1997cited by this paper
• Infants' Recognition of the Sound Patterns of Their Own Names

  1995cited by this paper
• Developmental changes in perception of nonnative vowel contrasts.

  1994cited by this paper
• Acoustic characteristics of American English vowels.

  1994cited by this paper
• Linguistic experience alters phonetic perception in infants by 6 months of age.

  1992cited by this paper
• The CHILDES project: tools for analyzing talk

  1992cited by this paper
• Cross-language speech perception: Evidence for perceptual reorganization during the first year of life

  1984cited by this paper
• Proceedings of the National Academy of Sciences

  1914cited by this paper
• PSYCHOLOGICAL SCIENCE Research Article Using Speakers ’ Referential Intentions to Model Early Cross-Situational Word Learning

  year unknowncited by this paper

• Simulating vocal learning of spoken language: Beyond imitation

  2023cites this paper
• Infant Phonetic Learning as Perceptual Space Learning: A Crosslinguistic Evaluation of Computational Models

  2023cites this paper
• Do Infants Really Learn Phonetic Categories?

  2021cites this paper
• Early phonetic learning without phonetic categories: Insights from large-scale simulations on realistic input

  2021cites this paper
• Measuring the perceptual availability of phonological features during language acquisition using unsupervised binary stochastic autoencoders

  2019cites this paper
• Training ASR Models By Generation of Contextual Information

  2019cites this paper
• Language input and outcome variation as a test of theory plausibility: The case of early phonological acquisition

  2018cites this paper
• Model-Free Context-Aware Word Composition

  2018cites this paper
• Visually Grounded Learning of Keyword Prediction from Untranscribed Speech

  2017cites this paper
• Unsupervised Learning of Semantic Audio Representations

  2017cites this paper
• BOOTSTRAPPING LANGUAGE ACQUISITION 2 1 . 1 Theoretical underpinnings Our model falls under the general umbrella of “ Semantic Bootstrapping ”

  2017cites this paper
• Bootstrapping language acquisition.

  2017cites this paper
• ABX-Discriminability Measures and Applications

  2016cites this paper
• Joint Word Segmentation and Phonetic Category Induction

  2016cites this paper
• Acquiring sounds and meaning jointly in early word learning

  2015cites this paper
• Unsupervised word discovery from speech using automatic segmentation into syllable-like units

  2015cites this paper
• A Rudimentary Lexicon and Semantics Help Bootstrap Phoneme Acquisition

  2014cites this paper
• Who “it” is influences what “it” does: Discourse effects on children’s syntactic parsing

  year unknowncites this paper