Modeling Different Aspects of Child Language Acquisition as a Probabilistic Process

Alishahi will first present a probabilistic model of word learning by children. A major source of disagreement among the different theories of word learning is whether children are equipped with special mechanisms and biases for word learning, or their general cognitive abilities are adequate for the task. Alishahi will present a novel computational model of early word learning which learns word meanings as probabilistic associations between words and semantic elements, using an incremental learning mechanism, and drawing only on general cognitive abilities. The computational simulations of the model demonstrate that much about word meanings can be learned from naturally-occurring child-directed utterances (paired with meaning representations), without using any special biases or constraints, and without any explicit developmental changes in the underlying learning mechanism. Furthermore, our model provides explanations for the occasionally contradictory child experimental data, and offers predictions for the behaviour of young word learners in novel situations.

Children use their knowledge of word meanings in order to learn the structural properties of the language. Alishahi will also present a probabilistic usage-based model of verb argument structure acquisition that can successfully learn abstract knowledge of language from instances of verb usage, and use this knowledge in various language tasks. The model further demonstrates the feasibility of a usage-based account of language learning, and provides concrete explanation for the observed patterns in child language acquisition.