Grammatical Error Correction

Course Description

Grammatical Error Correction (GEC) is a task of automatically detecting and correcting errors in different texts. It has a wide range of applications, from proof-reading to assist language learners and children with developmental language disorders.

This course is a project seminar for senior undergraduate students, and highlights learning by doing. In this course, we will show students how to build neural networks and guide them to master seq2seq framework.

Instructor: Meng Li

Time: Fridays, 10:15 - 16:00 (May 24, 31 and July 5, 12)

Room: May 24 - 2.14.0.18; May 31 - 2.10.0.26; July 5 - 2.05.0.04; July 12 - 2.14.0.18. (Different locations!😅)

Course Management System: Moodle

Syllabus

  • Day 1 (2024-05-24):
    • 10:15 - 12:00. From Perceptron to Feed-Forward Neural Networks
    • 13:00 - 14:45. Backpropagation and Learning Tricks
    • 15:00 - 16:00. PyTorch Tutorial
  • Day 2 (2024-05-31):
    • 10:15 - 12:00. RNN and Seq2seq
    • 13:00 - 14:45. A Primer on Grammatical Error Correction
    • 15:00 - 16:00. Implementing Baseline with PyTorch
  • Day 3 (2024-07-05):
    • 10:15 - 12:00. DL Projects in the Real World
    • 13:00 - 16:00. Project Troubleshooting and QA
  • Day 4 (2024-07-12):
    • 10:15 - 12:00. Project Presentation
    • 13:00 - 16:00. Project Presentation

Requirement

Prerequisites. You are expected to be proficient in Python programming.

Format. In the first two blocks, I will prepare the essential knowledge that you need to develop a project on grammatical error correction. In June, you are expected to implement your project individually. Your project should change some variables (architectures, representations, languages, etc.) compared with the baseline. In the third block, I will present how to improve DL projects, and also help to tackle practical issues. In the fourth block, each student should present their project and explain the code. I will ask random questions to check if you really understand your code.

Grading

This seminar belongs to “Methods of computational linguistics and subject hours/Methoden der Computerlinguistik und Ver-suchspersonenstunden (LIN-BS-070)”.

  • To complete this seminar, you are exptected to complete a software project on grammatical error correction, and give a presentation to explain your code at the end of this course.

  • (Optional) To pass the LIN-BS-070 module examination, you need to complete a term paper related to one of two seminars. If you would like to write a term paper on grammatical error correction, please talk to me.

Contact

Feel free to contact Meng at meng.li (at) uni-potsdam.de for any questions.