Skip to content

ThoughtTreasure commonsense knowledge base and architecture for natural language processing

License

Notifications You must be signed in to change notification settings

ameliamariecollins/thoughttreasure

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

thoughttreasure

ThoughtTreasure commonsense knowledge base and architecture for natural language processing

http://www.amazon.com/gp/product/1478171650/

Functionality

Comprehensive natural language processing/commonsense platform
Question answering
Chatterbot
Story understanding
Classified ad parsing
Simulation of a 2-dimensional virtual world
Identification of entities in text
Information extraction from ASCII tables
Concordance generation
Shell (command interpreter)
Web API
Java-based client API
Python-based client API
Knowledge base flat files

Applications

Commonsense reasoning
Question answering
Information extraction
Interactive fiction

Source Code

Language: ANSI C
Lines of code including comments: 75,000

Knowledge base and lexicon

Upper ontology
Domain lower ontologies (clothing, food, music, and so on)
Concepts, words, and phrases defined using concise language
English words and phrases: 35,000
French words and phrases: 20,000
Closed-class lexical items not in WordNet (adverbs and adverbials, prepositions, determiners, pronouns, conjunctions, interjections, temporal relations, and so on)
Concepts: 25,000
Assertions: 50,000 including hierarchical links
Scripts: 100 with events and roles

Linguistic features

English and French lexicon/ontology
Verb/nominalization argument structure
Name dictionary with grouping of related names
Text agents: part-of-speech tagging, identification of names, places, products, dates, phone numbers, email headers
English and French syntactic parser: base component, filters, transformations
English and French semantic parser: intension/extension, relative clauses, appositives, genitives, tense/aspect
Anaphoric parser: deixis, determiners, pronouns, c-command
Understanding agents: converting surfacy parse into detailed understanding, steering planning agents, contexts, emotions, goals, question answering, asking clarifying questions, appointments, sleep, grids by analogy
English and French generator
Learning of new words using derivational rules
Learning of new inflections by algorithmic/analogical morphology

Simulation features

Space represented by 2-dimensional grids connected by wormholes
Prototypical grids: house, apartment, restaurant, theater, street, subway
Planning agents for simulating human behavior: graspers, containers, ptrans++, atrans++, mtrans++, interpersonal relations
Planning agents for simulating device behavior: telephone, television

Procedures

Database: Assertion/retrieval of assertions
Contexts
Anagram finder
Palindrome finder
Inverted dictionary generator
English/French faux ami finder
Intension resolver: find objects matching a description
Subgoaling
2-dimensional (occupancy array) path planner
Grid operations (distance, subspace, etc.)
Intergrid path planner
Trip planner
Clothing color matching
Operations for parts and wholes of objects
Operations for nested space (room, floor, building, city, planet)
Operations for large space (planetary distance, polity containment)
Theorem prover
Temporal projection
Abduction
Relation learner (corporate successions)
Assertion learner
Language translation
Dictionary generator
Free association generator
Corpus analysis tools
Multi-language, multi-stream discourse channel
Chatterbot main loop
Typing simulator with errors
HTML utilities
Report generation

About

ThoughtTreasure commonsense knowledge base and architecture for natural language processing

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • C 76.4%
  • HTML 18.0%
  • Java 2.3%
  • Python 2.2%
  • CSS 0.4%
  • Makefile 0.2%
  • Other 0.5%