Please enable JavaScript.
Coggle requires JavaScript to display documents.
Natural Language Processing (NLP) - Coggle Diagram
Natural Language Processing (NLP)
Definition of NLP
Interactions between computers and human languages
Computational techniques for analyzing and representing text
Requirements for NLP
Linguistics
Phonology
Morphology
Phonetics
Lexicon
Cognitive Psychology
Computer Science
What is Language
Formal vs. Natural (Rules vs. Ambiguity
Syntax: Analyzing Sentence Structure
Semantics: Understanding Sentence Meaning
Discourse Analysis: Text as a Whole (Aims, Speech Acts)
Pragmatics: Extra Meaning (World Knowledge, Intentions)
Challenge: Natural Language Understanding
History
1950s: Turing Test, Georgetown Experiment
1970s: Ontologies, Chatterbots
1960s: Eliza (Chatbot), "Blocks World"
1980s: Shift to Machine Learning
Machine Learning
Statistical Models
Inferring Rules from Data
AI Capability
NLP is an "AI-Complete" Problem
Solving it would create Human-Level AI
Challenges
Phonology: Speech Recognition, Part-of-Speech Tagging
Morphology: Segmentation, Morphemes & Words
Lexicon: Dictionary, Word Sense Disambiguation, Named Entity Recognition
syntax - generative grammar, co-ref resolution, parsing, auto-summary, xml, relationship extraction
discourse analysis - topic segmentation and recognition
pragmatics - common sense, world knowledge, machine translation (all) natural language generation, natural language understanding, question answering
N-gram character model
N-Letter Sequences: Uses fixed-length sequences of characters.
Markov Chain Probability: Next character depends only on the preceding
Prediction: Makes predictions based on a given corpus of text.
Training: Can be trained on text data to improve accuracy.
Common sense Logic
Common sense logic involves everyday reasoning that mimics the human ability to make judgments based on typical experiences. It understands context and applies general knowledge to interpret situations. This approach allows for practical decision-making by applying intuitive solutions to problems and adjusting to new information similarly to human thought processes.
Web Links
https://www.scienceabc.com/social-science/what-is-common-sense.html
https://www.britannica.com/topic/logic
Logical Variations
Logical variations refer to different methods of reasoning and problem-solving. Deductive logic derives specific conclusions from general principles, while inductive logic infers general principles from specific observations. Abductive logic proposes the most likely explanation based on available evidence. Non-classical logic includes variations like fuzzy logic, which deals with reasoning that is approximate rather than fixed and exact.
Introduction
- Replicating human thinking involves creating systems that mimic human reasoning, decision-making, and problem-solving. These systems aim to understand context, apply common sense, and adapt to new information. The goal is to achieve human-like intelligence in machines.