ECE594: Mathematical Models of Language (Spring 2021)

Course Details

Department of Electrical and Computer Engineering
University of Illinois at Urbana-Champaign
Instructor: Prof. Suma Bhat

Lectures (All US Central Time):

• 9:30 AM - 10:50 PM TR

Course Zoom link

Gradescope entry code 74ENE4

Course Outline

Language technologies using AI and NLP are at work in our daily lives in tasks ranging from grammatical error correction to online language translation to online question answering. The intricacies of natural language we use as part of our everyday activities pose distinct challenges for computers that process unstructured text in the absence of real-word context or intent. In this course, we understand and analyze classical and recent computational models for addressing these challenges. The course will be centered around the following broad themes each over 5 weeks.

• Modeling language-related properties

• Applications of language processing

• Computational social science

We cover each of these themes in depth, discussing the core aspect of the theme via lectures and supplementing this with discussions using key papers on each topic. Students are expected to contribute to the discussion by weekly readings and presentations of the research papers.

Course Objectives

• To give the student a feel for the area of natural language processing by understanding the core challenges for processing language and current methods of cutting-edge research. 
• To enable the student to develop critical research skills, through literature review in the area, organize and share ideas via oral and written presentations, as well as providing constructive feedback to peers.

Prerequisites

• Basic Probability and Statistics​ 

• Foundations of Machine Learning (e.g. CS 446 or equivalent mastery of the material)

  We will rely on several machine learning concepts, including formulating cost functions, taking derivatives and performing optimization with gradient descent. If you already have basic machine learning and/or deep learning knowledge, the course will be easier; however you could still take this course if you can work your way through concurrent efforts. There are several excellent introductions to ML, in webpage, book, and video form. Some of them are listed in the Reference Books section.