Skip to content Skip to navigation

Natural Language Processing

16:198:533

An in-depth study of the ideas and techniques underlying the development of a computational theory of human language use, which is necessary both for practical interactive dialogue applications, as well as for the scientific study of human communication.

The course is intended for computer science graduate students, as well as students in allied areas (information science, linguistics, philosophy, psychology, computer engineering) who have interests in artificial intelligence and computational linguistics.

Category: 
B
Prerequisite: 

Either 198:530 or permission of the instructor.

Topics: 

This course surveys the models and reasoning required in computational systems that use natural language to communicate. The modeling problem is to take our theories and observations of human language use, and refine them into a formal description suitable for computation. The reasoning problems involve drawing inferences from these models to support effective communication between agents, despite differences in knowledge and ability, and despite ambiguity and noise in messages. In exploring these problems, we will cover the following topics:
    
1. Linguistic description and computational models: syntax, semantics, discourse. 
2. Communicative intent and real-world knowledge, beliefs, and plans. 
3. Conversational structure: modeling human behavior and expectations in a dialogue.
4. Reasoning with linguistic models: parsing, generation and dialogue management. 
5. Selected topics such as intelligent information retrieval, cognitive modeling, machine translation, and acquisition of linguistic knowledge.

Expected Work: 

Regular readings and homework assignments; a final project consisting of the student's choice of a term paper or a programming project.

Exams: 
None
Professor: 
Matthew Stone
Semester: 
Spring
Course Type: 
Graduate

Check the University Schedule of Classes to see if this course is open.

Request a Special Permission Number here if the class is full.