Combining competing language understanding approaches in an intelligent tutoring system

Pamela W. Jordan, Maxim Makatchev, Kurt VanLehn

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

When implementing a tutoring system that attempts a deep understanding of students' natural language explanations, there are three basic approaches to choose between; symbolic, in which sentence strings are parsed using a lexicon and grammar; statistical, in which a corpus is used to train a text classifier; and hybrid, in which rich, symbolically produced features supplement statistical training. Because each type of approach requires different amounts of domain knowledge preparation and provides different quality output for the same input, we describe a method for heuristically combining multiple natural language understanding approaches in an attempt to use each to its best advantage. We explore two basic models for combining approaches in the context of a tutoring system; one where heuristics select the first satisficing representation and another in which heuristics select the highest ranked representation.

Original languageEnglish (US)
Pages (from-to)346-357
Number of pages12
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3220
StatePublished - 2004
Externally publishedYes

Fingerprint

Intelligent Tutoring Systems
Intelligent systems
Natural Language
Classifiers
Language
Heuristics
Students
Domain Knowledge
Grammar
Preparation
Choose
Strings
Classifier
Output
Model
Context
Corpus
Training
Text

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

@article{2d5224e54f674008aefb5dfd430cedfb,
title = "Combining competing language understanding approaches in an intelligent tutoring system",
abstract = "When implementing a tutoring system that attempts a deep understanding of students' natural language explanations, there are three basic approaches to choose between; symbolic, in which sentence strings are parsed using a lexicon and grammar; statistical, in which a corpus is used to train a text classifier; and hybrid, in which rich, symbolically produced features supplement statistical training. Because each type of approach requires different amounts of domain knowledge preparation and provides different quality output for the same input, we describe a method for heuristically combining multiple natural language understanding approaches in an attempt to use each to its best advantage. We explore two basic models for combining approaches in the context of a tutoring system; one where heuristics select the first satisficing representation and another in which heuristics select the highest ranked representation.",
author = "Jordan, {Pamela W.} and Maxim Makatchev and Kurt VanLehn",
year = "2004",
language = "English (US)",
volume = "3220",
pages = "346--357",
journal = "Lecture Notes in Computer Science",
issn = "0302-9743",
publisher = "Springer Verlag",

}

TY - JOUR

T1 - Combining competing language understanding approaches in an intelligent tutoring system

AU - Jordan, Pamela W.

AU - Makatchev, Maxim

AU - VanLehn, Kurt

PY - 2004

Y1 - 2004

N2 - When implementing a tutoring system that attempts a deep understanding of students' natural language explanations, there are three basic approaches to choose between; symbolic, in which sentence strings are parsed using a lexicon and grammar; statistical, in which a corpus is used to train a text classifier; and hybrid, in which rich, symbolically produced features supplement statistical training. Because each type of approach requires different amounts of domain knowledge preparation and provides different quality output for the same input, we describe a method for heuristically combining multiple natural language understanding approaches in an attempt to use each to its best advantage. We explore two basic models for combining approaches in the context of a tutoring system; one where heuristics select the first satisficing representation and another in which heuristics select the highest ranked representation.

AB - When implementing a tutoring system that attempts a deep understanding of students' natural language explanations, there are three basic approaches to choose between; symbolic, in which sentence strings are parsed using a lexicon and grammar; statistical, in which a corpus is used to train a text classifier; and hybrid, in which rich, symbolically produced features supplement statistical training. Because each type of approach requires different amounts of domain knowledge preparation and provides different quality output for the same input, we describe a method for heuristically combining multiple natural language understanding approaches in an attempt to use each to its best advantage. We explore two basic models for combining approaches in the context of a tutoring system; one where heuristics select the first satisficing representation and another in which heuristics select the highest ranked representation.

UR - http://www.scopus.com/inward/record.url?scp=35048843281&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35048843281&partnerID=8YFLogxK

M3 - Article

VL - 3220

SP - 346

EP - 357

JO - Lecture Notes in Computer Science

JF - Lecture Notes in Computer Science

SN - 0302-9743

ER -