Abstract
This paper uses an unsupervised model of grounded language acquisition to study the role that social cues play in language acquisition. The input to the model consists of (orthographically transcribed) child-directed utterances accompanied by the set of objects present in the non-linguistic context. Each object is annotated by social cues, indicating e.g., whether the caregiver is looking at or touching the object. We show how to model the task of inferring which objects are being talked about (and which words refer to which objects) as standard grammatical inference, and describe PCFG-based unigram models and adaptor grammar-based collocation models for the task. Exploiting social cues improves the performance of all models. Our models learn the relative importance of each social cue jointly with word-object mappings and collocation structure, consistent with the idea that children could discover the importance of particular social information sources during word learning.
Original language | English |
---|---|
Title of host publication | Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics, ACL 2012 |
Place of Publication | Stroudsburg, PA |
Publisher | Association for Computational Linguistics (ACL) |
Pages | 883-891 |
Number of pages | 9 |
Volume | 1 |
ISBN (Print) | 9781937284244 |
Publication status | Published - 2012 |
Event | 50th Annual Meeting of the Association for Computational Linguistics, ACL - 2012 - Jeju Island, Korea, Republic of Duration: 8 Jul 2012 → 14 Jul 2012 |
Other
Other | 50th Annual Meeting of the Association for Computational Linguistics, ACL - 2012 |
---|---|
Country/Territory | Korea, Republic of |
City | Jeju Island |
Period | 8/07/12 → 14/07/12 |