How to Generate Cloze Questions from Definitions: a Syntactic Approach Donna Gates, Gregory Aist (Iowa State University), Jack Mostow, Margaret McKeown.

Slides:



Advertisements
Similar presentations
Cloze Measure What is Measured? Comprehension We are only able to infer what others comprehend All measures of comprehension are indirect Students read.
Advertisements

Teaching Writing for the Russian State Exam
Test Taking Strategies
Reliability and Validity of Researcher-Made Surveys.
Computational language: week 10 Lexical Knowledge Representation concluded Syntax-based computational language Sentence structure: syntax Context free.
1 Verbal Ability Vocabulary Building and Sentence Completion.
Chapter 4 Syntax.
Carnegie Mellon Generating Questions Automatically from Informational Text Wei Chen, Gregory Aist, and Jack Mostow Project LISTEN, School of Computer Science.
Grammars, Languages and Parse Trees. Language Let V be an alphabet or vocabulary V* is set of all strings over V A language L is a subset of V*, i.e.,
Analysing Syntax 1 Lesson 8B.
Mining Data from Randomized Within-Subject Experiments in an Automated Reading Tutor Joseph E. Beck and Jack Mostow Project LISTEN (
Detecting Prosody Improvement in Oral Rereading Minh Duong and Jack Mostow Project LISTEN Carnegie Mellon University The research.
LING 364: Introduction to Formal Semantics Lecture 4 January 24th.
Language processing What are the components of language, and how do we process them?
Lessons from generating, scoring, and analyzing questions in a Reading Tutor for children Jack Mostow Project LISTEN (
Building Background F9: Vocabulary Bell Work: Write 3 strategies you can use in the classroom to help students link past learning with new concepts Opening.
THE FUNDAMENTALS OF WRITING ONE CAN ONLY READ WHEN SOMEONE HAS WRITTEN SOMETHING.
KS1 Information Sessions for Parents
ACT Reading.
Contraction Memory Contents: Box, set of contraction cards, instructions, 6 answer keys Instructions: 1.Place all cards blank side down in a group on a.
Welcome to Gymnasium Horn Upper School! ©2009 Dr. Claudia Börger Greetings in formal contexts  Hello, I just wanted to introduce myself. I’m...
AN ARTICLE. WHAT IS IT? a piece of writing (in a newspaper or magazine) goal: to make an issue interesting to readers could be: descriptive, narrative,
Chapter 10: Compilers and Language Translation Invitation to Computer Science, Java Version, Third Edition.
Learning to Transform Natural to Formal Language Presented by Ping Zhang Rohit J. Kate, Yuk Wah Wong, and Raymond J. Mooney.
The BIG question: How is this going to work? Planning-my part Patience- all of our part Must Do’s-the student’s part.
Dear Judge 9B Module. Activity One: Quickwrite It’s not a secret that addictions—to drugs, alcohol, smoking, and more—are common among teens. The question.
Unit 7: M A S E D I.
 Definition of a quality Audit  Types of audit  Qualifications of quality auditors  The audit process.
Youth speak out about issues. Schedule  May 5th - May 9th (generating ideas/planning/researching)  May 13th - May 16th (drafting)  May 19th - May 23rd.
Improving Search Results Quality by Customizing Summary Lengths Michael Kaisser ★, Marti Hearst  and John B. Lowe ★ University of Edinburgh,  UC Berkeley,
Department of Software and Computing Systems Research Group of Language Processing and Information Systems The DLSIUAES Team’s Participation in the TAC.
Effective Claims A checklist for writing strong claims.
Carnegie Mellon Mostow 12/7/2015, p. 1 The Sounds of Silence: Towards Automated Evaluation of Student Learning in a Reading Tutor that Listens Jack Mostow.
Writing a formal letter
Carolyn Penstein Rosé Language Technologies Institute Human-Computer Interaction Institute School of Computer Science With funding from the National Science.
NLP. Introduction to NLP Background –From the early ‘90s –Developed at the University of Pennsylvania –(Marcus, Santorini, and Marcinkiewicz 1993) Size.
Paper 5: Speaking Part 2: photo description and
Carnegie Mellon How does the amount of context in which words are practiced affect fluency growth? Experimental results Jack Mostow, Jessica Nelson, Martin.
Judging Tips for Elementary Projects 2012 Scott Ferguson, Ph.D. Atlantic Turf & Ornamental Consulting.
Academic Vocabulary Unit 7 Cite: To give evidence for or justification of an argument or statement.
An Introduction to Semantic Parts of Speech Rajat Kumar Mohanty rkm[AT]cse[DOT]iitb[DOT]ac[DOT]in Centre for Indian Language Technology Department of Computer.
3.3 A More Detailed Look At Transformations Inversion (revised): Move Infl to C. Do Insertion: Insert interrogative do into an empty.
October 10, 2003BLTS Kickoff Meeting1 Transfer with Strong Decoding Learning Module Transfer Rules {PP,4894} ;;Score: PP::PP [NP POSTP] -> [PREP.
NATURAL LANGUAGE PROCESSING
VOCABULARY FOR INFORMATIVE PAPER THE WORDS YOU NEED.
Facilitators: The Encore Team.  Today we will address: Manners How should you look?
New Bulgarian University ENGLISH B2 FINAL/EXAM TEST FORMAT, CONTENT and PROCEDURES © Angel Rundaliev, NBU, Sofia, Bulgaria.
Reading Strategies We Use Every Day. 1. Creating Mental Images Good readers:  Visualize and create pictures in their mind  Organize details in a “mental.
Key Stage 2 Portfolio. Llafaredd / Oracy Darllen / Reading Ysgrifennu / Writing Welsh Second Language.
Strategies & Knowledge for Independent Learning individual Work SKIL SKIL cards are sets of cards that explain how to use different learning strategies.
-what you should include for the different questions. Section A- there is NO choice of question in this section- it will focus on either Vietnam or the.
Useful sentences 1. 尽管他们说不同种英语,英语本族人可以 理解对方. Native English speakers can understand each other even if they don’t speak the same kind of English. 2.
Writing an Opinion Piece. Expressing an Opinion An opinion is a thought or belief about someone or something.
Wei Chen, Jack Mostow Gregory Aist Project LISTEN
Minimal English Test vs. TOEIC®
What is Political Science and Why Do We Compare?
Module 9 Population.
Discussion and Conclusion
Detecting Prosody Improvement in Oral Rereading
Year 2: How to help your child
Donna M. Gates Carnegie Mellon University
Independent versus Computer-Guided Oral Reading:
An Embedded Experiment to Evaluate the Effectiveness of Vocabulary Previews in an Automated Reading Tutor Jack Mostow, Joe Beck, Juliet Bey, Andrew Cuneo,
Chapter 2 What speakers know.
FCE (FIRST CERTIFICATE IN ENGLISH) General information.
Key Ideas: unlock What is it about? How do I know?
This material is based upon work supported by the National Science Foundation under Grant #XXXXXX. Any opinions, findings, and conclusions or recommendations.
Word of the Day List nd 9 Weeks.
Using Phonemic Awareness &
WELCOME TO STRUCTURES CONGRESS 2020
Presentation transcript:

How to Generate Cloze Questions from Definitions: a Syntactic Approach Donna Gates, Gregory Aist (Iowa State University), Jack Mostow, Margaret McKeown (University of Pittsburgh), and Juliet Bey Project LISTEN ( School of Computer Science, Carnegie Mellon Universitywww.cs.cmu.edu/~listen Conclusions The research reported here was supported by the Institute of Education Sciences, U.S. Department of Education, through Grant R305A The opinions expressed are those of the authors and do not necessarily represent the views of the Institute and the U.S. Department of Education. We thank Joe Valerie for evaluating our data. Quality was surprisingly good given our vocabulary expert’s high standards. Humans are costly, but hand-filtering yielded more examples per minute than writing them by hand. Need to reduce hand-filtering time to scale up from several hundred target words to several thousand. Future work: filter out possible answers as distracters by measuring their semantic similarity to the correct answer. Definitions Written Specifically for 2 nd and 3 rd Graders 23 Syntactic Phrase PatternsExample NP a large flat piece of meat (AP (RB? ADJ)) they are not polite (NP it) ADVP? (VP? (VP …)) it smells or tastes very bad (NP you) ADVP? (VP? (VP …) you leave them and never go back If you abandon someone or something, you _____ and never go back. ask about them leave them If you abandon someone or something, you _____ and never go back. leave them believe in it If you abandon someone or something, you leave them and never go back. A steak is a large flat piece of meat. If something is crude, it is not made well. If someone is crude, they are not polite. Parse definition Find definition content Match syntactic phrase patterns Collect syntactic phrases with same pattern in other definitions Generate cloze question form (ROOT (S (SBAR (IN If) (S (NP (PRP you)) (VP (VBP abandon) (NP (NN something) (CC or) (NN someone))))) (,,) (NP (PRP you)) (VP (VP (VBP leave) (NP (PRP them))) (CC and) (ADVP (RB never)) (VP (VBP go) (ADVP (RB back)))) (..))) you VP|know them pretty well|familiar.a you VP|know them|recognize.v you VP|lean slowly from one side to the other|sway.v you VP|leave them|abandon.v you VP|like them very much|attract.v you VP|like them very much|fond.a you VP|look everywhere for them very carefully|scour.v you VP|look for it|seek.v you VP|make a hole in it|pierce.v If you abandon someone or something, you _____ and never go back. A steak is _____. If something is crude, it is _____. If someone is crude, they are _____. Choose possible distracters with same pattern Filter distracters automatically, then manually Automatic and Manual Filters If you abandon something or someone, you _____ and never go back. leave them (answer) look for it (good distracter) like them very much (good distracter) look for someone that you have not met before (too long) do it (too vague: could fit) are very mean (too close/specific: could fit if you think it is mean to abandon something) lose it (too close in meaning) According to our vocabulary expert, an acceptable distracter: 1.Should be same syntactic phrase type as the answer (e.g., VP vs. NP vs. AP vs. PP) 2. Should have roughly the same length as the answer 3. Should not be a possible answer 4.Should be grammatical when inserted into the blank (loose subject-verb-agreement ok) Evaluating Generated Output Definition Cloze Questions Development set: 300+ definitions Evaluation set: 33 definitions excluded from development Generated output: 91 cloze questions and 522 non-unique distracters Two (human) judges: 1.disagreed on 14 cloze forms (which words to delete) 2.for the 77 cloze forms agreed upon, a.accepted 100% of cloze sentences b.accepted 71% of distracters c.disagreed on 94 distracters 3.human filtering yielded 7.19 acceptable items/minute 4.hand-writing examples yielded 1.29 items/minute Reading Tutor gives 2 nd and 3 rd graders vocabulary instruction on selected words in the story a child is reading, plus additional practice activities on later days. To remind the student what a word means, we generate a simple two-choice cloze question from its definition.