CS 5/662, Winter 2021

Natural Language Processing

Overview

This course covers key algorithms and modeling techniques for processing human language sequences, which are needed for applications such as Automatic Speech Recognition and Machine Translation. Both statistical and symbolic approaches to modeling natural language phonology, morphology, and syntax are presented, along with widely used algorithms for efficiently learning and applying different kinds of natural language grammars. There is an emphasis on algorithms and data structures that scale up to handle very large real-world data sets, such as newswire text. The course includes several challenging hands-on programming assignments. Suggested prerequisite: CSE 560 or equivalent. Python programming experience is highly recommended, as is familiarity with regular expressions.

Note that due to the coronavirus pandemic, this term, our class sessions and labs will be held via Webex. The call-in information can be found on the main course page in Sakai. If synchronous distance learning becomes an issue for you this term (e.g. due to computer or internet access issues, childcare scheduling conflicts, etc.), please contact me as soon as possible so we can discuss alternatives.

Textbooks & Resources

• Jurafsky & Martin, Speech and Language Processing, 3rd edition (revision in progress; draft chapters posted online)
• Goldberg, Neural Network Methods for Natural Language Processing
• Eisenstein, Introduction to Natural Language Processing
  • Note that a draft version of the book is available online; as far as I can tell, it’s quite similar to the final published version.
• Bender, Linguistic fundamentals for natural language processing : 100 essentials from morphology and syntax

Two of the textbooks we will be using are available electronically from the OHSU Library, and both of the others are available in draft form online.

If possible, I do encourage you to buy a copy of at least the Eisenstein book, so as to support the publishing of high-quality NLP texts.

Course goals

By the end of the course, students will:

• Be conversant with a variety of both traditional statistical and modern neural methods in natural language processing;
• Have gained practical, hands-on experience with computational tools and techniques for the above;
• Be able to empirically evaluate NLP systems;
• Be familiar and conversant with issues of bias and transparency in NLP models and applications

Logistics

When & Where?

Wednesdays & Fridays, 16:00 – 17:30 via Webex

Starts: January 6

Ends: March 19

Who?

Instructor: Steven Bedrick

Office Location (not that it matters): Gaines Hall, 21

Office Hours: Mondays 16:00 – 18:00; By appointment otherwise

---

Page last updated: 2022-12-21 14:05