Compiler Theory

• Standard Compilers work on Multi-Pass system
• Tokenization (regular languages, words)
• Syntax tree generation (context free languages)
• Optimization
• Code Generation
• A. Aho, R. Sethi and J. Ullman (1986) Compilers; Principles, Techniques and Tools. Addison Wessley
• The first two of these are applicable to Natural Language Processing (NLP).
• Of course people work on one pass. When you hear a sentence or read book, you largely do all the processing as you go.