Please enable JavaScript.
Coggle requires JavaScript to display documents.
Natural Language Processing (NLP) - Coggle Diagram
Natural Language Processing (NLP)
Introduction
Language & Linguistics
History of NLP
Challenges: Common sense understanding
Formal vs Natural Language
Formal languages: Python, Java (rules, clear meaning)
Natural language: Human languages (ambiguity, evolving)
Key Fields Needed
Linguistics (structure of language)
Computer Science (internal data representation)
Cognitive Psychology (language and human cognition)
Linguistic Concepts
Phonetics & Phonology (speech sounds)
Morphology (word structure)
Lexicon (word meanings and relations)
Syntax (grammar structures)
Semantics (meaning)
Discourse Analysis (connections across sentences)
Pragmatics (context, world knowledge)
History of NLP
1950s: Turing Test, Georgetown Experiment (early translation)
1960s: Eliza chatbot, SHRDLU (blocks world)
1970s: Conceptual ontologies, chatterbots (ALICE, Jabberwacky)
1980s+: Move from rule-based to Machine Learning
Machine Learning for NLP
Statistical models
Use of big data
Focus on probabilities, not hard-coded rules
Current Progress
Mostly Solved:
OCR
Named Entity Recognition (NER)
POS Tagging
Good Progress:
Sentiment Analysis
Co-reference Resolution
Word Sense Disambiguation (WSD)
Parsing & Information Extraction
Machine Translation
Still Hard:
Natural Language Understanding
Logical Reasoning
Dialogue Systems
Common Sense and Logic
Humans rely on hidden, implicit knowledge
Example challenges: river crossing puzzle, birthday surprise planning
Efforts: ConceptNet, Open Mind Common Sense databases