Identifying words to explain to a reader: A preliminary study

Gregory Aist
Proceedings of the Seventeenth National Conference on Artificial Intelligence (AAAI-2000), August, 2000, pp. 1061.


Abstract
The core idea of this paper is familiar to teachers: While a child is reading, explain unfamiliar words. Project LISTEN's Reading Tutor (A HREF="http://www.cs.cmu.edu/~listen">http://www.cs.cmu.edu/~listen) listens to children read aloud and helps them learn to read. We want the Reading Tutor to explain unfamiliar words.

To elicit explanations from an expert, the computer should suggest -- or let the expert select -- words to annotate.

To capture explanations, the expert will type in and then narrate an explanation. Text and narration will be saved for later use (Mostow & Aist 1999).

To utilize explanations during assisted reading, we will display the explanations as extra sentences to be read aloud with the computer's help. Explanations will be provided on student request or computer tutor initiative.

We focus here on how to select words for annotation. How? Annotating all words may be too expensive in preparation time or storage space. Kids may not know which words they need help on. Grade-level wordlists may not include all the hard words in a story. Perhaps adult readers can identify words to explain.[1]


Notes
Associated Lab(s) / Group(s): Project LISTEN
Associated Project(s): Project LISTEN\'s Reading Tutor

Text Reference
Gregory Aist, "Identifying words to explain to a reader: A preliminary study," Proceedings of the Seventeenth National Conference on Artificial Intelligence (AAAI-2000), August, 2000, pp. 1061.

BibTeX Reference
@inproceedings{Aist_2000_3673,
   author = "Gregory Aist",
   title = "Identifying words to explain to a reader: A preliminary study",
   booktitle = "Proceedings of the Seventeenth National Conference on Artificial Intelligence (AAAI-2000)",
   pages = "1061",
   month = "August",
   year = "2000",
}