Modeling the contribution of phonotactic cues to the problem of word segmentation

Abstract

How do infants find the words in the speech stream? Computational models help us understand this feat by revealing the advantages and disadvantages of different strategies that infants might use. Here, we outline a computational model of word segmentation that aims both to incorporate cues proposed by language acquisition researchers and to establish the contributions different cues can make to word segmentation. We present experimental results from modified versions of Venkataraman's (2001) segmentation model that examine the utility of: (1) language-universal phonotactic cues; (2) language-specific phonotactic cues which must be learned while segmenting utterances; and (3) their combination. We show that the language-specific cue improves segmentation performance overall, but the language-universal phonotactic cue does not, and that their combination results in the most improvement. Not only does this suggest that language-specific constraints can be learned simultaneously with speech segmentation, but it is also consistent with experimental research that shows that there are multiple phonotactic cues helpful to segmentation (e.g. Mattys, Jusczyk, Luce & Morgan, 1999; Mattys & Jusczyk, 2001). This result also compares favorably to other segmentation models (e.g. Brent, 1999; Fleck, 2008; Goldwater, 2007; Johnson & Goldwater, 2009; Venkataraman, 2001) and has implications for how infants learn to segment.

Online Learning Mechanisms for Bayesian Models of Word Segmentation

“…Similarly, phonotactic constraints differ between languages, with legal words in one language being illegal in others. Thus, one would normally assume that in order to use these cues, infants must have already learned some of the words in the language in order to identify the dominant stress patterns and phonotactics [though see Blanchard et al (2010) for one way language-specific phonotactics might be learned at the same time the initial segmentation problem is being solved]. Since the point of word segmentation is to identify words in the first place, needing to know words in order to learn segmentation cues creates a chicken-and-egg problem.…”

“…Using a sliding window of a fixed length, two "voting experts" use accumulated entropy knowledge to vote whether a boundary should be inserted between two phonemes; if the number of votes exceeds a pre-determined threshold, the learner inserts a word boundary. Blanchard et al (2010) created PHOCUS (PHOnotactic CUe Segmenter), an online learner that couples statistical word learning with phonotactic constraints. In particular, Blanchard et al demonstrate a way in which language-specific phonotactic constraints, realized as likely and unlikely phonemic sequences, can be learned at the same time that segmentation is being attempted and an explicit lexicon is built.…”

“…Instead, they trained over the entire Bernstein corpus and were evaluated by how well they segmented that corpus. In addition, PHOCUS performance was computed after discarding the first 1000 utterances [see Blanchard et al (2010) for discussion as to why]. The results for the transitional probability learner are from an implementation that computes transitional probabilities based on an initial sweep through the corpus (technically making it a batch algorithm; online versions are possible but would require some form of smoothing to avoid zero-probability transitions).…”

Learning novel phonotactics from exposure to continuous speech

“…Computational work has shown that phonotactic constraints are learnable from transcriptions of continuous speech, and that such constraints have a positive effect on word segmentation (Adriaans & Kager, 2010;Blanchard et al, 2010;Brent & Cartwright, 1996;Cairns et al, 1997;Daland & Pierrehumbert, 2011). Phonotactic constraints might thus arise before the lexicon is in place, in particular from co-occurrence probabilities in continuous speech.…”

Utility-based evaluation metrics for models of language acquisition: A look at speech segmentation

“…Fortunately, experimental evidence suggests that infants can leverage language-independent probabilistic cues to identify that initial seed pool of words (Saffran, Aslin, & Newport, 1996;Aslin, Saffran, & Newport, 1998;Thiessen & Saffran, 2003;Pelucchi, Hay, & Saffran, 2009). This had led to significant interest in the early probabilistic segmentation strategies infants use (Brent, 1999;Batchelder, 2002;Goldwater et al, 2009;Blanchard, Heinz, & Golinkoff, 2010;Pearl et al, 2011;Lignos, 2011).…”