Natural language processing (NLP) uses computers to interpret and manipulate human language. Because natural language is ambiguous, inconsistent, and constantly changing, practical systems must handle more than fixed grammar rules. section 14.1
A practical NLP workflow begins with manageable text tasks:
- Preprocess text input into numeric input suitable for computation. section 14.0
- Build simple text classification models as a foundation for more complicated architectures. section 14.0
- Account for context and generalization rather than relying only on hand-written language rules. section 14.1
From hand-coded translation to modern applications
An early Russian-to-English system used six grammar rules and a small lookup table to translate 60 selected sentences. Its success on curated examples contrasted with the difficulty of building a robust general translator. Later NLP advances supported chatbots, programming assistants, transcription, and improved machine translation. section 14.1 section 1.9
