MasterŐs Program in Computational Linguistics

 

USC offers two programs in the area of Computational Linguistics (also called Natural Language Processing and Human Language Technology):

-         This program offers an MS degree.  It is centered in the Department of Linguistics and focuses on issues in Linguistics.  Its faculty are primarily experts in Linguistics. 

-         The other program (click here) offers MS and PhD degrees in Computer Science with an emphasis on Human Language Technology / Computational Linguistics, and focuses on all aspects of computational linguistics.  Its faculty are primarily experts in Computer Science, and are members of the world-renowned Natural Language research groups at the Information Sciences Institute (ISI) and Institute for Creative Technology (ICT). 

 

 

Faculty

Joseph Aoun, Professor, Linguistics, USC

Michael Arbib, Professor, Computer Science, USC

Robert Belvin, Computer Scientist, HRL Laboratories; Lecturer, Linguistics, USC

Bonnie Glover Stalls, Adjunct Professor, Linguistics, USC

Shrikanth Narayanan, Assistant Professor, Electrical Engineering, USC

Jean-Roger Vergnaud, Professor, Linguistics, USC

 

 

Research Areas

 

The faculty members comprise a group of internationally renowned scholars from Linguistics. Their areas of research help to shape research in the program. Within the field of Linguistics, prevailing models of formal grammar have been shaped to a great extent by these scholars whose expertise encompasses not only grammatical theory but a wide range of languages and language families. Jean-Roger Vergnaud established the notion of Abstract Case theory, which has remained one of the central concerns of formal grammatical theory for more than two decades. Joseph Aoun pioneered the study of Generalized Binding and has remained in the forefront of investigations of application of formal grammatical models to East Asian and Semitic languages. Bonnie Glover Stalls is working on automated information extraction for web-based resume text processing. She has developed Arabic lexical, syntactic, and semantic resources for broad-coverage Arabic-to-English machine translation system at ISI and has extensive industrial experience in multilingual appliations including voice-to-voice translation. n Computer Science Michael Arbib is active in computational and cognitive neuroscience as well as neuroinformatics. A topic of current interest relates the mechanisms of control of hand movements in monkeys to a new scenario for the origin of human language.  Shrikanth Narayanan is a member of the Electrical Engineering Department after working at AT&T Research. He focuses on automated speech recognition and speech synthesis.

 

 

Courses Offered

 

Required Courses

 

LING 530 Syntax (3 units, Fall) An intensive introduction to the principle and methods of grammatical analysis. The structure of sample English clauses will be discussed, and the essential formal notions and empirical results will be presented. In particular, the central concept of multiplicity of representation will be developed in its various forms. The concept is inseparable from that of lexical representation of rule. Properties of relations between concomitant representations, as well as a taxonomy of such relations, are discussed and exemplified. The central grammatical relations are formalized and illustrated. A range of central phenomena will be discussed and analyzed, with appeal to cross-linguistic evidence and considerations.

LING 534 Logic and the Theory of Meaning (3 units, Spring) An introduction to logic in preparation for advanced work in semantics and linguistic theory. The language of first-order logic, and introduction to truth-theoretic semantics for formal and natural languages. Compositionality. Predication, and reference. Syntax and semantics of quantification. Generalized quantifiers and their lexical properties. Proper Nouns, definite descriptions and descriptive anaphora.

LING 585 Computational Linguistics (3 units, Fall) This course presents an overview of computational systems that process natural language, in particular examining the role of linguistic knowledge and the procedures that implement it in working systems. Topics covered include speech recognition and generation, computational lexicography, morphological analysis, natural language parsing and its relation to syntactic theory, lexical classes and lexical semantics, and computational ontology. The course includes both hands-on and research components and emphasizes the expanding role linguistics can play in this emerging field, as well as how computational tools and techniques can contribute to linguistic research and theory.

 

Recommended Prerequisite Course

 

LING 500 Structure of Language (3 units, Fall) Development of analytical skills in syntax and semantics, with major attention to language universals and language typologies and their relevance to theories of language acquisition.

 

Breadth Requirement Courses

LING 533 Language Universals and Typology (3 units) Introduction to language universals and typology.

LING 538 Selected Topics in Romance Syntax (3 units) Overview of selected topics in Romance Syntax within a comparative perspective and their contribution towards the understanding of a general theory of grammar.  Prerequisite: departmental approval.

LING 539 Japanese/Korean Syntax and Theoretical Implications (3 units) Critical discussion of selected papers and dissertations on Japanese/Korean syntax and consideration of their theoretical implications.   Prerequisite: departmental approval.

LING 548 Lexical Semantics (3 units) The primary focus of this course is on how meaning is constituted within the linguistic unit of the word. Several proposals for Lexical Decomposition are examined with both theory-internal and cross-linguistic empirical evidence being considered. Lexicalization Patterns form a major part of this evidence (that is, different ways in which languages group meaning elements together to form words). Other important topics include the question of how the meaning of a word (especially verbs) is related to its syntactic properties (mapping hypotheses), which in turn brings in the topics of Thematic/Semantic Roles, Selectional features and Verb Classes.

LING 576 Psycholinguistics (3 units, Fall) Theories of acquisition; sentence and discourse processing; language and thought.   Prerequisite: departmental approval.

 

Relevant Courses in Other Departments

 

EE 519 Digital Speech Processing (3 units, Fall) Graduate introductory course on speech processing and speech recognition.

EE 599 Advanced Topics in Speech Recognition and Spoken Language Engineering (3 units, Spring)

 

CSCI 561a Artificial Intelligence (3 units, Fall or Spring) Foundations of symbolic intelligent systems. Agents, search, problem solving, representation, reasoning and symbolic programming.  Prerequisite: CSCI455x. Since this is in Computer Science, this course requires permission from the instructors.

CSCI 544 Natural Language Processing (3 units, Spring) This course covers the basic techniques of processing human language by computer. These include morphological analysis, parsing, semantic interpretation, and generation. We examine symbolic algorithms as well as some ways of acquiring linguistic knowledge automatically through statistical analysis. Techniques are presented in the context of applications like machine translation, text summarization, and information retrieval.  Pre or co-requisite: CSCI 561a.  Since this is in Computer Science, this course requires permission from the instructors. 

CSCI 562 Empirical Methods in Natural Language (3 units, Fall) Acquiring computer-tractable linguistic knowledge has always been a bottleneck in building automatic translation, speech recognizers, summarizers, grammar checkers, and information management systems. Some of this knowledge can now be statistically extracted from large texts. We will examine the state-of-the-art in statistical modeling, supervised training, and bootstrapping. The approach will be experimental; software tools will allow students to build their own applications and measure performance. Prerequisite: CSCI 561a. Since this is in Computer Science, this course requires permission from the instructors.

 

 

Application information

 

Who should apply:

-         Graduates in linguistics or related fields, or college seniors, interested in seeking a career in computational linguistics.

-         Professional linguists interested in computational methods.

-         Students looking to build a research-oriented resume before entering industry or a PhD program.

 

Admissions requirements:

Admissions to the Master's Program in Computational Linguistics requires:

-         a completed bachelor's degree (or equivalent) in linguistics, mathematics, or a related field from an accredited institution

-         a GPA of 3.0 or higher

-         satisfactory GRE test scores

-         -TOEFL test, for international students

-         the ability to program, with expertise in such computer languages as LISP, C++, PROLOG, PERL, or JAVA

-         proficiency in basic linguistics (phonology, morphology, syntax) and expertise in linguistics data analysis.

-         advanced knowledge or the equivalent of at least two years of study at the college level of a human language other than English.

 

Applicants also need letters of evaluation from at least three professors or co-workers and a one or two-page statement of purpose. A major or work experience in a closely related field, such as Linguistics, Computer Science, or a particular language, is helpful but not required.

 

Students interested in the program who have not fulfilled the linguistics requirements (basic phonology, morphology, and syntax, and linguistic data analysis) or the recommended language requirement are strongly encouraged to fulfill them before entering the program. To help strengthen students' backgrounds in these areas, it may be possible to take a prerequisite course in Computer Science (Introduction to Programming Systems Design) and one in Linguistics (Structure of Language). Students should be aware, however, that the breadth requirements or elective courses may require additional prerequisites.

 

Admissions procedure:

Prospective students should apply to the Office of Graduate Admissions for the Master's Program in Computational Linguistics (through the Department of Linguistics). Please submit the following:

1)      The original application and a $55.00 fee. Your application will not be processed until the fee is paid.

2)      GRE scores (must be sent to USC by the testing agency).

3)      One official transcript from all academic institutions and universities attended, and a certified English translation if the transcript is from an international institution.

 

You must also submit the following documents to the Department of Linguistics:

1)      A copy of the original application

2)      A copy of your GRE scores

3)      Copies of all transcripts as well as English translation of courses and degree,

4)      A detailed Statement of Purpose showing clarity of expression and organization, and specific information about interests and goals

5)      3 academic letters of recommendation, and

6)      A research paper, if possible

 

Please note that new students are admitted to the program in the Fall semester only and that we do not accept any documents by email.

 

 

Contact

 

Bonnie Glover Stalls

Adjunct Professor

Department of Linguistics

University of Southern California

Los Angeles, CA 90089-1693

Tel: +1-213-740-2986

Fax: +1-213-740-9306

email: bstalls@isearch.com