Mixed-initiative Control for Teaching and Learning in Disciple George Mason University School of Information Technology and Engineering Computer Science.

Slides:



Advertisements
Similar presentations
Learning Agents Center George Mason University Symposium on Reasoning and Learning in Cognitive Systems Stanford, CA, May 2004 Gheorghe Tecuci with.
Advertisements

Language Technologies Reality and Promise in AKT Yorick Wilks and Fabio Ciravegna Department of Computer Science, University of Sheffield.
Modelling with expert systems. Expert systems Modelling with expert systems Coaching modelling with expert systems Advantages and limitations of modelling.
Artificial Intelligence of a Scrabble System
Chapter 12 User Interface Design
Firewall Query Engine and Firewall Comparison Engine Mohamed Gouda Alex X. Liu Computer Science Department The University of Texas at Austin.
Supporting Business Decisions Expert Systems. Expert system definition Possible working definition of an expert system: –“A computer system with a knowledge.
Alternate Software Development Methodologies
Intelligent Agent for Designing Steel Skeleton Structures of Tall Buildings Zbigniew Skolicki Rafal Kicinger.
Actionable Intelligence: Continuous Discovery of Evidence, Hypotheses, and Arguments from Masses of Data A computational theory, methodology and system.
Assurance through Enhanced Design Methodology Orlando, FL 5 December 2012 Nirav Davé SRI International This effort is sponsored by the Defense Advanced.
An Agent Capable of Learning to Create and Maintain Websites Anthony Tomasic, Ravi Mosur Alex Rudnicky, Raj Reddy, John Zimmerman Carnegie Mellon University.
11.1 Lecture 11 CASE tools IMS Systems Design and Implementation.
Chapter 11 Artificial Intelligence and Expert Systems.
1 8. Safe Query Languages Safe program – its semantics can be at least partially computed on any valid database input. Safety is tied to program verification,
Spoken Dialogue Technology How can Jerry Springer contribute to Computer Science Research Projects?
Interact: RETSINA’s Agent Editor Provides a GUI interface to agent’s task and reduction libraries Allows –quick development of new libraries –easy extensions.
These courseware materials are to be used in conjunction with Software Engineering: A Practitioner’s Approach, 6/e and are provided with permission by.
These courseware materials are to be used in conjunction with Software Engineering: A Practitioner’s Approach, 6/e and are provided with permission by.
Expert Systems Infsy 540 Dr. Ocker. Expert Systems n computer systems which try to mimic human expertise n produce a decision that does not require judgment.
What is Business Analysis Planning & Monitoring?
21 September 2005 Gheorghe Tecuci Learning Agents Center and Computer Science Department School of Information Technology and Engineering George Mason.
Beyond Intelligent Interfaces: Exploring, Analyzing, and Creating Success Models of Cooperative Problem Solving Gerhard Fischer & Brent Reeves.
Concept Visualization for Ontologies of Artificial Intelligence Yu Suo TJHSST Computer Systems Lab George Mason University.
System Design: Designing the User Interface Dr. Dania Bilal IS582 Spring 2009.
1 Learning Agents Laboratory Computer Science Department George Mason University Prof. Gheorghe Tecuci 1. Introduction.
 2004, G.Tecuci, Learning Agents Center CS 785 Fall 2004 Learning Agents Center and Computer Science Department George Mason University Gheorghe Tecuci.
11 C H A P T E R Artificial Intelligence and Expert Systems.
 2003, G.Tecuci, Learning Agents Laboratory 1 Learning Agents Laboratory Computer Science Department George Mason University Prof. Gheorghe Tecuci 11.
Introduction Complex and large SW. SW crises Expensive HW. Custom SW. Batch execution Structured programming Product SW.
1 Learning Agents Laboratory Computer Science Department George Mason University Prof. Gheorghe Tecuci 2. Projects and assignments.
Learning Agents Center George Mason University Computer Science Department Partners Day Symposium May 4, 2004 Gheorghe Tecuci, Mihai Boicu, Dorin Marcu,
 G.Tecuci, Learning Agents Laboratory Learning Agents Laboratory Department of Computer Science George Mason University George Tecuci
How Solvable Is Intelligence? A brief introduction to AI Dr. Richard Fox Department of Computer Science Northern Kentucky University.
 2004, G.Tecuci, Learning Agents Center 1 CS 7850 Fall 2004 Learning Agents Center and Computer Science Department George Mason University Gheorghe Tecuci.
 2003, G.Tecuci, Learning Agents Laboratory 1 Learning Agents Laboratory Computer Science Department George Mason University Prof. Gheorghe Tecuci 1.
Chap. 5 Building Valid, Credible, and Appropriately Detailed Simulation Models.
1 COSC 4406 Software Engineering COSC 4406 Software Engineering Haibin Zhu, Ph.D. Dept. of Computer Science and mathematics, Nipissing University, 100.
Learning Agents Center George Mason University Symposium on Reasoning and Learning in Cognitive Systems Stanford, CA, May 2004 Gheorghe Tecuci with.
SCIENCE A system of knowledge about the natural world and the methods used to find that knowledge.
 2004, G.Tecuci, Learning Agents Center 1 CS 7850 Fall 2004 Learning Agents Center and Computer Science Department George Mason University Gheorghe Tecuci.
 2004, G.Tecuci, Learning Agents Center CS 785 Fall 2004 Learning Agents Center and Computer Science Department George Mason University Gheorghe Tecuci.
Distributed Models for Decision Support Jose Cuena & Sascha Ossowski Pesented by: Gal Moshitch & Rica Gonen.
1 USC INFORMATION SCIENCES INSTITUTE EXPECT TEMPLE: TEMPLate Extension Through Knowledge Acquisition Yolanda Gil Jim Blythe Information Sciences Institute.
David W. Aha Navy Center for Applied Research in AI Naval Research Laboratory Co-chairs Gheorghe Tecuci Learning Agents Center George Mason University.
PnP Networks Self-Aware Networks Self-Aware Networks Self-Healing and Self-Defense via Aware and Vigilant Networks PnP Networks, Inc. August, 2002.
Constructivism: The Social Construction of International Politics POL 3080 Approaches to IR.
Search Engine Optimization © HiTech Institute. All rights reserved. Slide 1 Click to edit Master title style What is Business Analysis Body of Knowledge?
Of An Expert System.  Introduction  What is AI?  Intelligent in Human & Machine? What is Expert System? How are Expert System used? Elements of ES.
Unclassified//For Official Use Only 1 RAPID: Representation and Analysis of Probabilistic Intelligence Data Carnegie Mellon University PI : Prof. Jaime.
ITEC 1010 Information and Organizations Chapter V Expert Systems.
What Is Science?. 1. Science is limited to studying only the natural world. 2. The natural world are those phenomena that can be investigated, discovered,
Design Evaluation Overview Introduction Model for Interface Design Evaluation Types of Evaluation –Conceptual Design –Usability –Learning Outcome.
Managing Qualitative Knowledge in Software Architecture Assesment Jilles van Gurp & Jan Bosch Högskolan Karlskrona/Ronneby in Sweden Department of Software.
Analysis and Reporting Toolset (A&RT): Lessons on how to develop a system with an external partner David Smith AstraZeneca.
Gheorghe Tecuci 1,2, Mihai Boicu 1, Dorin Marcu 1 1 Learning Agents Laboratory, George Mason University 2 Center for Strategic Leadership, US Army War.
 2004, G.Tecuci, Learning Agents Center CS 785 Fall 2004 Learning Agents Center and Computer Science Department George Mason University Gheorghe Tecuci.
Intelligent Systems Development
Decision Support Systems
Improving Undergraduate STEM Education: Education and Human Resources (IUSE: EHR) Teaching Critical Thinking Skills in Science with sInvestigator: Project.
Knowledge Acquisition and Problem Solving
Intro to Expert Systems Paula Matuszek CSC 8750, Fall, 2004
Agent Development Project
Evidence Evaluation Year 1 Focus
Sample elicitation scripts Execution of the elicitation scripts
Unit 1 Scientific Inquiry
Agent Development Project
Multistrategy Rule Refinement
Expert Knowledge Based Systems
Architecture Issue in the New Disciple System
Presentation transcript:

Mixed-initiative Control for Teaching and Learning in Disciple George Mason University School of Information Technology and Engineering Computer Science Department Learning Agents Laboratory Acapulco, Mexico - August 9 th, 2003 Mihai Boicu, Gheorghe Tecuci, Dorin Marcu, Cristina Boicu, Bogdan Stanescu IJCAI – 2003 Workshop on Mixed-initiative Intelligent Systems

© 2003 Learning Agents Laboratory – Mihai Boicu Slide: 2 Disciple Approach Learning Agent Interface Learning Problem Solving Ontology + Rules The expert teaches the agent in a way that resembles how the expert would teach a person. The agent learns from the expert, building, verifying and improving its knowledge base Disciple is an evolving theory, methodology and family of agent shells for rapid development and use of knowledge bases and agents, by subject matter experts, with limited assistance from knowledge engineers The agent help the expert to teach The expert help the agent to learn Mixed-initiative

© 2003 Learning Agents Laboratory – Mihai Boicu Slide: 3 Teaching Methodology

© 2003 Learning Agents Laboratory – Mihai Boicu Slide: 4 I need to Therefore Test whether the US_Sectret_Service_1943 that protects President_Roosevelt has any significant vulnerability Test whether President_Roosevelt has mean_to_be_protected What is a means to protect President_Roosevelt from all threats? President_Roosevelt is protected by the US_Secret_Service_1943. IF Test whether a controlling element has a critical requirement The controlling element is ?O1 The critical requirement is ?O2 THEN: Test whether a protection agency that protects a controlling element has any significant vulnerability The controlling element is ?O1 The protection agency is ?O3 Explanation ?O1 is_protected_by ?O3 ?O3 provides ?O2 ?O2 is requirement_to_be_protected ?O3 is protection_agency Plausible Upper Bound Condition ?O1 is agent is_protected_by ?O3 ?O2 is requirement_to_be_ protected ?O3 is protection_agency provides ?O2 IF Test whether ?O1 has ?O2 Question: What is a means to protect ?O1 from all threats? Answer: ?O1 is protected by the ?O3. Plausible Lower Bound Condition ?O1 is head_of_government is_protected_by ?O3 ?O2 is requirement_to_be_ protected ?O3 is personal_protection_ agency provides ?O2 THEN Test whether the ?O3 that protects ?O1 has any significant vulnerability means_to_ be_protected Rule President_ Roosevelt is_protected_by critical_requirement_ for_a_capability strategic_COG_ relevant_factor agent object person controlling_leader head_of_government political_leader Example US_Secret_ Service_1943 protection_ agency personal_ protection_agency provides requirement_to_ be_protected critical_requirement_for_ a_capability_of_a_leader critical_requirement Knowledge Representation

© 2003 Learning Agents Laboratory – Mihai Boicu Slide: 5 Mixed-initiative Problem Solving and Learning Input Task Generated Reduction Mixed-Initiative Problem Solving Ontology + Rules Reject Reduction Accept Reduction New Reduction Rule Refinement Task Refinement Rule Refinement Modeling Formalization Learning Solution

© 2003 Learning Agents Laboratory – Mihai Boicu Slide: 6 Multi-agent architecture of Teaching Assistant Modeling Assistant Distributed GUI Word Completion Agent Example Editor Agent Example Completion Agent Mixed-initiative Control Agent Teaching Assistant Rule Learning Assistant Problem Solving Assistant Teaching Assistant GUI Modeling Assistant GUI Example Analyzer Agent GUI Example Editor Agent GUI Example Completion Agent GUI Example Analyzer Agent Task Formalization Assistant Rule Analyzer Agent Implicit Explanations Agent Explanation Generation Agent Problem Solving Agent Word Completion Agent GUI Rule Refinement Assistant Rule Regeneration Agent Example-based Ontology Learning Exception-based Ontology Learning

© 2003 Learning Agents Laboratory – Mihai Boicu Slide: 7 Mixed-initiative Interaction during Modeling Distributed GUI Example Editor Agent User typed letter “i” in the question of the example Update the screen appearance Notify registered processes Word Completion Agent Update example question (example-id, new-question, version) If the same version update GUI appearance What is a means to protect Pr| Example Editor President_Roosevelt President Word completion Automatically identify known terms Modify the internal structure of the example Notify registered processes If the same version, validate component and show completion suggestions Example Completion Agent User modified example question (example-id, new-question, version) Compute the most plausible word completion, if any Invalidate component selection Update the preemptive computation of the next element (the answer) Update completions (word-completions, version)

© 2003 Learning Agents Laboratory – Mihai Boicu Slide: 8 Evaluation of the Modeling of Expert’s Reasoning Experimental results for the modeling adviser 2001  The expert succeeded to model more complex reasoning steps in the 2002 experiment than in the 2001 one.  The experts assessed that the suggestions provided by the modeling adviser were understandable and useful  The experts assessed that the modeling adviser is relatively easy to use and well organized. 2002

© 2003 Learning Agents Laboratory – Mihai Boicu Slide: 9 Evaluation of the Rule Analyzer Rule Analyzer After a rule is learned (or refined) the agent analyze it to identify potential problems and to suggest corrections:  Variable flow problems (some input variables are not linked to output variables)  Under-constrained rule (the rule generates too many solutions during problem solving)... Experimental results  The experts have always responded to the warnings of the Rule Analyzer  The experts have assessed that the suggestions of the Rule Analyzer were generally understandable and useful

© 2003 Learning Agents Laboratory – Mihai Boicu Slide: 10 Acknowledgements This research was performed in the Learning Agents Laboratory (LALAB) of George Mason University under the direction of Dr. Gheorghe Tecuci. Research of the LALAB is sponsored by the Defense Advanced Research Projects Agency, the Air Force Office of Scientific Research, Air Force Research Laboratory, Air Force Material Command, US Air Force, and the US Army, under agreement number F and grant number F The members of the LALAB contributed to the success of the developed systems.