Automated grammatical error detection for language learners /
It has been estimated that over a billion people are using or learning English as a second or foreign language, and the numbers are growing not only for English but for other languages as well. These language learners provide a burgeoning market for tools that help identify and correct learners'...
Main Authors: | , , , |
---|---|
Format: | Book |
Language: | English |
Published: |
Cham, Switzerland :
Springer,
[2014]
|
Edition: | Second edition |
Series: | Synthesis lectures on human language technologies ;
#25 |
Subjects: |
Table of Contents:
- 1. Introduction
- 1.1 Introduction to the second edition
- 1.2 New to the second edition
- 1.3 Working definition of grammatical error
- 1.4 Prominence of research on English language learners
- 1.5 Some terminology
- 1.6 Automated grammatical error detection: NLP and CALL
- 1.7 Intended audience
- 1.8 Outline
- Background
- Special problems of language learners
- Evaluating error detection systems
- Data-driven approaches to articles and prepositions
- Collocation errors
- Different errors and different approaches
- Annotating learner errors
- Emerging directions
- 10. Conclusion
- A. Appendix A. Learner corpora
- Bibliography
- Authors' biographies
- 2. Background
- 2.1 In the beginning
- 2.2 Introduction to data-driven and hybrid approaches
- 3. Special problems of language learners
- 3.1 Errors made by English language learners
- 3.2 The influence of L1
- 3.3 Challenges for English language learners
- 3.3.1 The English preposition system
- 3.3.2 The English article system
- 3.3.3 English collocations
- 3.4 Summary
- 4. Evaluating error detection systems
- 4.1 Traditional evaluation measures
- 4.2 Evaluation measures for shared tasks
- 4.3 Evaluation using a corpus of correct usage
- 4.4 Evaluation on learner writing
- 4.4.1 Verifying results on learner writing
- 4.4.2 Evaluation on fully annotated learner corpora
- 4.4.3 Using multiple annotators and crowdsourcing for evaluation
- 4.5 Statistical significance testing
- 4.6 Checklist for consistent reporting of system results
- 4.7 Summary
- 5. Data-driven approaches to articles and prepositions
- 5.1 Extracting features from training data
- 5.2 Types of training data
- 5.2.1 Training on well-formed text
- 5.2.2 Artificial errors
- 5.2.3 Error-annotated learner corpora
- 5.2.4 Comparing training paradigms
- 5.3 Methods
- 5.3.1 Classification
- 5.3.2 N-gram statistics, language models, and web counts
- 5.3.3 Web-based methods
- 5.4 Two end-to-end systems: criterion and MSR ESL assistant
- 5.5 Summary
- 6. Collocation errors
- 6.1 Defining collocations
- 6.2 Measuring the strength of association between words
- 6.3 Systems for detecting and correcting collocation errors
- 7. Different errors and different approaches
- 7.1 Heuristic rule-based approaches
- 7.1.1 Criterion system
- 7.1.2 ESL assistant
- 7.1.3 Other heuristic rule-based approaches
- 7.2 More complex verb form errors
- 7.3 Spelling errors
- 7.4 Punctuation errors
- 7.5 Detection of ungrammatical sentences
- 7.6 Summary
- 8. Annotating learner errors
- 8.1 Issues with learner error annotation
- 8.1.1 Number of annotators
- 8.1.2 Annotation schemes
- 8.1.3 How to correct an error
- 8.1.4 Annotation approaches
- 8.1.5 Annotation tools
- 8.2 Annotation schemes
- 8.2.1 Examples of comprehensive annotation schemes
- 8.2.2 Example of a targeted annotation scheme
- 8.3 Proposals for efficient annotation
- 8.3.1 Sampling approach with multiple annotators
- 8.3.2 Crowdsourcing annotations
- 8.3.3 Mining online community-driven revision logs
- 8.4 Summary
- 9. Emerging directions
- 9.1 Shared tasks in grammatical error correction
- 9.1.1 The 2011 HOO task
- 9.1.2 The 2012 HOO task
- 9.1.3 The CoNLL 2013 shared task
- 9.1.4 Summary
- 9.2 Machine translation and error correction
- 9.2.1 Noisy channel model
- 9.2.2 Round trip machine translation (RTMT)
- 9.3 Real-time crowdsourcing of grammatical error correction
- 9.4 Does automated error feedback improve writing?