Knowledge Management Challenges for Question Answering Vinay K. Chaudhri SRI International White Paper Co-authors: Ken Barker (UT), Tom Garvey (SRI), Ken.

Slides:



Advertisements
Similar presentations
Mitsunori Ogihara Center for Computational Science
Advertisements

FT228/4 Knowledge Based Decision Support Systems Knowledge Engineering Ref: Artificial Intelligence A Guide to Intelligent Systems, Michael Negnevitsky.
DARPA SHAKEN Virus-Invades-Cell Invade VirusCell Attach Penetrate Release Move invader thing invaded barrier Cell-membrane has-part subevent penetrator.
Proceedings of the Conference on Intelligent Text Processing and Computational Linguistics (CICLing-2007) Learning for Semantic Parsing Advisor: Hsin-His.
Common Core State Standards (CCSS) Nevada Joint Union High School District Nevada Union High School September 23, 2013 Louise Johnson, Ed.D. Superintendent.
Chapter 6: Design of Expert Systems
Quality evaluation and improvement for Internal Audit
AQUAINT Pilot evaluation for knowledge-oriented systems April 20, 2005.
AFROSAI-E COOPERATION WITH WGITA African Organisation of English-speaking Supreme Audit Institutions.
Semantic Web Technologies Lecture # 2 Faculty of Computer Science, IBA.
Unit 2: Engineering Design Process
1 Artificial Intelligence Applications Institute Centre for Intelligent Systems and their Applications Stuart Aitken Artificial Intelligence Applications.
Additional Unit 2 Lecture Notes New Instructional Design Focus School of Education Additional Unit 2 Lecture Notes New Instructional Design Focus School.
CHATS IN THE CLASSROOM: EVALUATIONS FROM THE PERSPECTIVES OF STUDENTS AND TUTORS AT CHEMNITZ UNIVERSITY OF TECHNOLOGY, COMMUNICATION ON TECHNOLOGY AND.
CLC reading program Nguyen Thi Thu Trang. In-class activities Assignment Assessment Add your text in here Reading program Objectives Contents.
CLOSE READING & ANNOTATING WHAT IT IS AND HOW TO DO IT.
IXA 1234 : C++ PROGRAMMING CHAPTER 1. PROGRAMMING LANGUAGE Programming language is a computer program that can solve certain problem / task Keyword: Computer.
Research Methodology and Writing 2013 Fall. The Outline Form P.45 P.45.
The E ngineering Design Process Foundations of Technology The E ngineering Design Process © 2013 International Technology and Engineering Educators Association,
Design Process TED 105 Fall Define the Problem (ask) Clearly state the problem –Problem statement –Needs assessment –Design criteria & goals.
©2003 Paula Matuszek CSC 9010: Text Mining Applications Dr. Paula Matuszek (610)
1 Report Writing Report writing. 2 Contents What is a report? Why write reports? What makes a good report? Fundamentals & methodology »Preparation »Outlining.
1.  Interpretation refers to the task of drawing inferences from the collected facts after an analytical and/or experimental study.  The task of interpretation.
Notes for Week 11 Term project evaluation and tips 3 lectures before Final exam Discussion questions for this week.
IT’S ALIVE!!. or is it??? Are viruses alive ? –How can you tell if something is alive? –What would your evidence be or look like?
Chapter 9.  Multimedia- communication that involves more than one format.  Basic Functions- can incorporate text, graphics, pictures and photos, video,
Module 4—Literacy Strands Arts Education. Learning Outcomes Participants will: explore the relationship between the new Essential Standards and the Common.
Welcome: Language Arts 8 Literature Circles Independent Novel LA8U7L1.
Mgt 540 Intro 1 Mgt 540 Research Methods. Mgt 540 Intro 2 Introduction Emeric Solymossy –Pronounced “Shoi-moshi” “Dr. E ” Availability / Accessibility.
Knowledge Support for Modeling and Simulation Michal Ševčenko Czech Technical University in Prague.
Decision Support Systems سيستم ‌ هاي تصميم ‌ يار Lecturer: A. Rabiee Rabiee.iauda.ac.ir.
HHS 307 Week 3 Final Paper Outline and Annotated Bibliography Check this A+ tutorial guideline at
Foundations of Technology The Engineering Design Process
GAT Preparation - the written component
Engineering Fundamentals and Problem Solving, 6e
The Reverse Engineering Process
Fundamentals of Information Systems, Sixth Edition
CSC 221: Computer Programming I Spring 2010
Guangbing Yang Presentation for Xerox Docushare Symposium in 2011
IB Assessments CRITERION!!!.
Knowledge Evolution Tools
CSC 221: Computer Programming I Fall 2005
Reports Chapter 17 © Pearson 2012.
Lecture #11: Ontology Engineering Dr. Bhavani Thuraisingham
Architecture Components
Chapter 6: Design of Expert Systems
Progress from Urban Planning Department on Work Supported by CTE
Fundamentals of Excel WORKSHOP DESCRIPTION Register Here
Illustrations of different approaches Peter Clark and John Thompson
FUNDAMENTAL ISSUES IN CULTURE TESTING
Social Knowledge Mining
Lecture 1: Course Outline and Introduction
The Starting Point: Asking Questions
Data Warehousing and Data Mining
Critical Thinking Angela Mazzetti
CSE 635 Multimedia Information Retrieval
Text-to-Text Generation
Foundations of Technology The Engineering Design Process
Writing Learning Outcomes
Foundations of Technology The Engineering Design Process
Capstone Team Project title
Advanced Design Applications The Engineering Design Process
Teamwork is crucial to success in an organization
AFROSAI-E COOPeRATION WITH WGITA
Knowledge Theory & A Unified Theory of AI
CS565: Intelligent Systems and Interfaces
How Should You Participate in this Course?
CONSTRUCTIVE ALIGNMENT
AICE General Paper What IS this class?.
Introduction to information retrieval
Presentation transcript:

Knowledge Management Challenges for Question Answering Vinay K. Chaudhri SRI International White Paper Co-authors: Ken Barker (UT), Tom Garvey (SRI), Ken Murray (SRI), Bruce Porter (UT), Tomas Uribe (SRI)

Outline  Interest in QA  Similarities and differences to AQUAINT  Comments on AQUAINT

Interest in QA  Co-chaired AAAI Symposia Question Answering Systems (Fall’99) Mining Answers from Text and KBs (Spring 2002)

Interest in QA  Associated with DARPA projects High Performance Knowledge Bases Knowledge bases by knowledge engineers Rapid Knowledge Formation Knowledge bases by domain expert

Rapid Knowledge Formation Piece of knowledge Knowledge Entry Tool Knowledge Base Solve Problems

Rapid Knowledge Formation Piece of knowledge Knowledge Entry Tool Knowledge Base Solve Problems Text Answer questions

Rapid Knowledge Formation Piece of knowledge Knowledge Entry Tool Knowledge Base Solve Problems Text Answer questions Biology textbook Questions at the back of the book

Rapid Knowledge Formation Piece of knowledge Knowledge Entry Tool Knowledge Base Solve Problems Text Answer questions Biology textbook Questions at the back of the book We never work with text while answering questions

Rapid Knowledge Formation Piece of knowledge Knowledge Entry Tool Knowledge Base Solve Problems In the head of a human - Course of Action - Critiquing knowledge Simulation

Similarities and Differences  Similarities Knowledge can come from text QA as an evaluation technique  Differences Reducing the cost of hand crafted KB construction Logical representations for masses

Comments on AQUAINT  Breadth  Synergy  Evaluation

Comments on AQUAINT Breadth  Impressed by breadth Free Text Media Language Structured Data

Comments on AQUAINT Breadth  Impressed by breadth  Large amount of text is fundamental Free Text Multi-media Multi-lingua Structured Data

Comments on AQUAINT Synergy  Numerous synergies by combining free text with structure Handcrafted KBs (Cyc, Wordnet, Framenet,..) Learning from Text (ISI, LCC,..) Annotations (Time markup language – NERC) More is possible …

Comments on AQUAINT Evaluation  Opinion Questions Well-defined tasks Clearly articulated technical problem Clear statement of issues Well executed

Finally….  AQUAINT seems like a very well-managed program Friendly PM Executive Committee Strong participation from Intelligence Community

Thank You!

What Could Be Done?  Enhance Annotations  Knowledge Resources  Modality Independent Tools

Enhance Annotations  Look for places where manual annotation effort is already being invested XML documents Intelligence report summarization Corpus construction  Use RKF tools to do annotations Annotations come from a KB Not limited to keywords Can do inference with background knowledge

Knowledge Resources  SHAKEN incorporates significant general knowledge We give our system for free for research use  We can scale it to cover sizable text  Will be fun to combine it with large text/multi-media collections for fixed- domain QA

Modality Independent Tools  QA Management Managing interaction Managing the lifecycle of a question  Answer fusion  Source validity