A unified model of early word learning: Integrating statistical and social cues

@article{citeulike:2114082, abstract = {Previous studies on early language acquisition have shown that word meanings can be acquired by an associative procedure that maps perceptual experience onto linguistic labels based on cross-situational observation. Recently, a social-pragmatic account focuses on the effect of the child's social-cognitive capacities, such as joint attention and intention reading. This paper argues that statistical and social cues can be seamlessly integrated to facilitate early word learning. To support this idea, we first introduce a statistical learning mechanism that provides a formal account of cross-situational observation. A unified model is then presented that is able to make use of different kinds of embodied social cues, such as joint attention and prosody in maternal speech, in the statistical learning framework. In a computational analysis of infant data, our unified model performs significantly better than the purely statistical approach in computing word-meaning associations.}, author = {Yu, Chen and Ballard, Dana H. }, booktitle = {Selected papers from the 3rd International Conference on Development and Learning (ICDL 2004), 3rd International Conference on Development and Learning; Time series prediction competition: the CATS benchmark}, citeulike-article-id = {2114082}, doi = {10.1016/j.neucom.2006.01.034}, journal = {Neurocomputing}, keywords = {cross-situational, word-learning}, month = {August}, number = {13-15}, pages = {2149--2165}, posted-at = {2007-12-14 15:00:19}, priority = {4}, title = {A unified model of early word learning: Integrating statistical and social cues}, url = {http://dx.doi.org/10.1016/j.neucom.2006.01.034}, volume = {70}, year = {2007} }

TY - JOUR ID - citeulike:2114082 L3 - citeulike-article-id:2114082 TI - A unified model of early word learning: Integrating statistical and social cues T2 - Selected papers from the 3rd International Conference on Development and Learning (ICDL 2004), 3rd International Conference on Development and Learning; Time series prediction competition: the CATS benchmark JF - Neurocomputing VL - 70 IS - 13-15 SP - 2149 EP - 2165 N2 - Previous studies on early language acquisition have shown that word meanings can be acquired by an associative procedure that maps perceptual experience onto linguistic labels based on cross-situational observation. Recently, a social-pragmatic account focuses on the effect of the child's social-cognitive capacities, such as joint attention and intention reading. This paper argues that statistical and social cues can be seamlessly integrated to facilitate early word learning. To support this idea, we first introduce a statistical learning mechanism that provides a formal account of cross-situational observation. A unified model is then presented that is able to make use of different kinds of embodied social cues, such as joint attention and prosody in maternal speech, in the statistical learning framework. In a computational analysis of infant data, our unified model performs significantly better than the purely statistical approach in computing word-meaning associations. KW - cross-situational KW - word-learning AU - Yu, Chen AU - Ballard, Dana PY - 2007/08// UR - http://dx.doi.org/10.1016/j.neucom.2006.01.034 ER -