An%20Interactive%20Dialogue%20System%20for%20Knowledge%20Acquisition%20in%20Cyc - PowerPoint PPT Presentation

About This Presentation
Title:

An%20Interactive%20Dialogue%20System%20for%20Knowledge%20Acquisition%20in%20Cyc

Description:

codify & enter each piece of knowledge, by hand. CYC. Large effort to date. 19 realtime years ... JerryAllison ?EXISTING-OBJECT-TYPE ?MUSICAL-PERFORMANCE-ORGANIZATION) ... – PowerPoint PPT presentation

Number of Views:56
Avg rating:3.0/5.0
Slides: 18
Provided by: gmat
Learn more at: http://lalab.gmu.edu
Category:

less

Transcript and Presenter's Notes

Title: An%20Interactive%20Dialogue%20System%20for%20Knowledge%20Acquisition%20in%20Cyc


1
An Interactive Dialogue System for Knowledge
Acquisition in Cyc
  • Michael Witbrock
  • 2003-08-09

2
Cyc Large Scale Knowledge Base
gt3,000 Microtheories
gt1,600,000 Assertions
Cyc
gt200,000 Rules
gt30,000 Relations
gt120,000 Concepts
3
How did CYC get this far?
Large effort to date 19 realtime years
substantial investment
2003
1984
learning by discovery
learning via natural language
codify enter each piece of knowledge, by hand
CYC
4
Analogy Developer
5
Phrase Disambiguator
6
Unknown Term
(isa PeggySue Song-CW)
7
Precision Suggestor
8
Concept Refinement Interview
(playsInstrumentInMusicalGroup JerryAllison
?EXISTING-OBJECT-TYPE ?MUSICAL-PERFORMANCE-ORGANIZ
ATION)
9
Concept Refinement Interview Why I asked
10
Concept Refinement Interview Induction
(isa JerryAllison ?NATIONALITY)
11
Mixed-Initiative Dialogue
User Interaction Agenda
12
Progress in Knowledge Entry Efficiency
50 40 30 20 10
Assertions per Hour
Manual KE
Feasibility Study 2000-10
RKF Year 1 2001-08
RKF Year 2 2002-09
13
Where do we want to go next?
  • Never make the user wait
  • Support both rapid and diligent parsing.
  • Improve ambiguity resolution
  • Deferred resolution by ambiguity tolerance
  • Anaphor resolution
  • Focus-based resolution
  • Transparent, correctable eager resolution
  • Conversational goals

14
Design Goal
Interlocutor
History / Status
Text / Template Entry
Suggestions for Efficient Knowledge Entry
Graphical View of Added Knowledge
15
Parse Pipeline
Syntactic Parser
Semantic Parser
Reformulator
Underspecified Semantic Formula
Parse Tree
Natural Language
Assertible Semantic Formula
Knowledge Base
16
Advanced parsing and discourse modeling
  • Interaction with Cyc should be a conversation
    that the user controls. Cycs responses should
    be
  • Relevant Track thread and overall goals
  • Correctable Detect, examine, correct
  • Learned Do better next time
  • Move to declarative representation and use
    inference more
  • Parsing and resolution rules are more
    transparent, flexible, and learnable
  • Deferred and inter-sentential resolution with
    ambiguity tolerance
  • User/Cyc relationship as teacher/student (or
    student/teacher)

17
(No Transcript)
Write a Comment
User Comments (0)
About PowerShow.com