Computational Linguistics

Computational linguistics studies language from a computational perspective — modelling, processing, and generating natural language with computers.

Scope

It covers natural language processing, parsing, machine translation, speech processing, and statistical and neural models of language.

Core questions

• How can computers process and generate human language?
• How can linguistic structure be modelled computationally?
• How can language data be used to learn language models?
• How are speech and text understood automatically?

Key concepts

• Natural language processing
• Parsing
• Machine translation
• Statistical language models
• Speech recognition
• Corpora

Key theories

Statistical NLP

Manning and Schütze synthesized the statistical, data-driven approach to language processing.

Speech and language processing

Jurafsky and Martin unified linguistic and computational approaches across speech and text.

History

Computational linguistics moved from rule-based systems to statistical methods (Manning & Schütze; Jurafsky & Martin) and, more recently, neural and large language models, central to modern language technology.

Debates

Rule-based versus data-driven approaches

Whether language technology is best built from linguistic rules or learned from data.

Key figures

• Christopher Manning
• Hinrich Schütze
• Daniel Jurafsky
• James Martin

Seminal works

• manning-schutze-1999
• jurafsky-martin-2000

Frequently asked questions

What is natural language processing?

The computational techniques for analysing, understanding, and generating human language, the core of computational linguistics.