Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Dal web 2.0 all'E-L 3.0


Published on

Introduzione all'incontro su EL e KM (15/7/08)

Published in: Technology, Education
  • Be the first to comment

Dal web 2.0 all'E-L 3.0

  1. 1. Intelligent Learning Environment Concept
  2. 2. LMS – ILE Goals <ul><li>To create a platform able to support : </li></ul><ul><ul><li>Agents with “ profilable subject matter “ expert knowledge </li></ul></ul><ul><ul><li>Relief Tutors from 1° level “ Subject matter “ tutoring activity </li></ul></ul><ul><ul><li>Students / Work Groups (LE) with 1° level “ Subject matter “ tutoring </li></ul></ul><ul><ul><li>Interpret the “tutoring role” according to : </li></ul></ul><ul><ul><ul><li>Knowledge / Subject matter profile </li></ul></ul></ul><ul><ul><ul><li>Pedagogical profile in conjunction with the “learning entity (LE)” profile </li></ul></ul></ul><ul><ul><li>AI Engine input </li></ul></ul><ul><ul><ul><li>LMS tracking data </li></ul></ul></ul><ul><ul><ul><li>LSA , or any type of system interpretable, external input </li></ul></ul></ul><ul><ul><ul><li>Profiles </li></ul></ul></ul><ul><ul><ul><li>Manual intervention from Experts, Tutors etc. </li></ul></ul></ul>
  3. 3. LMS – ILE Design issues <ul><li>Static Design strives for a simple, elegant, efficient solution to a single situation. </li></ul><ul><li>Outside of that situation the design is useless </li></ul><ul><li>Adaptation strives to just survive in a constantly changing situation. </li></ul><ul><li>Adaptation is continuously making design decisions </li></ul>Requirements Design Implement Test Water Fall Design Process Impl Impl Adaptive Control Loop Implementation Loads Capacity Cost QoS Adaptive Control Policy Conditions Conditions
  4. 4. LMS – ILE Environment Global Architecture of the ILE Platform
  5. 5. LMS – ILE Environment <ul><li>OLAT Java based LMS performing functions like: </li></ul><ul><ul><li>Student , Teacher, Authors administration </li></ul></ul><ul><ul><li>Group administration </li></ul></ul><ul><ul><li>E-Course administration </li></ul></ul><ul><ul><li>General access control </li></ul></ul><ul><ul><li>Global community functions like Forum, Wiki, Blog </li></ul></ul><ul><ul><li>Extended tailored functions like e-Portfolio etc. </li></ul></ul><ul><li>AI coaching / tutoring Engine </li></ul><ul><ul><li>Learning entity Agent (real-time student / group profiler) Curriculum Subagent (generating and adapting learner curriculum) </li></ul></ul><ul><ul><li>Knowledgebase Agent (Subject matter related) Pedagogical profiler Coach / Tutoring Subagent Subject matter profiler </li></ul></ul><ul><ul><li>Presentation Agent (Dynamic session initiator) </li></ul></ul><ul><ul><li>Pedagogical Agent </li></ul></ul><ul><ul><li>Coaching/Tutoring administrator Agent </li></ul></ul><ul><li>Authoring Middleware </li></ul><ul><ul><li>AIML / UML / SOAP utilities </li></ul></ul><ul><ul><li>AI ILE Authoring tools </li></ul></ul><ul><ul><li>Global Knowledgebase management utilities </li></ul></ul><ul><li>Standalone e-Course development environment </li></ul><ul><li>AI simulation / Test utilities </li></ul>ILE Components
  6. 6. OLAT – ILE INTEGRATION ILE-AI Push/Pull Servlet ILE-AI LMS Communicator AI - Engine
  7. 7. LMS – ILE Environment Plugins Blackboard (Method for inter Agent Communication) Agent Publish (Push) Subscribe (Pull) Message Queue
  8. 8. LMS – ILE AI Framework component servic e BB Behavior effecter coordinator sensor component service servic e component library Agent BB Behavior effecter coordinator sensor component component service servic e library Agent Abstracted Environment System specific Variable Boundary Application domain specific Cougaar Agent Reference Model Framework Infrastructure
  9. 9. LMS – ILE AI Framework <ul><li>Symbolize domain knowledge </li></ul><ul><li>Continuously executing </li></ul><ul><li>Responding to dynamic environment </li></ul><ul><li>No dependencies on other Plugins </li></ul><ul><li>Publish all relevant information to the blackboard </li></ul>Plug-in Functionality Allocators Assessors Cougaar Agent Data User Interface Expanders
  10. 10. Architectural Mapping Sensor-Based Control Loop Model-Based Control Loop Cognitive Control Loop Model Policy Situation inference rules days to minutes secs to msecs Network Disk management plane data plane Sensor/ Activity Proxy Agents Real-time Optimizer Agents processing status coordination resource status coordination resource trends coordination Cognitive Learner Agents processing. trends coordination Situation Predictor Agents processing pattern coordination resource pattern coordination Sensor/Activity Proxy Agents Processing Units CPU
  11. 11. Architectural Mapping <ul><li> Application </li></ul><ul><li>Functional modules (oval shaped) </li></ul><ul><li>Underlying distributed environment </li></ul><ul><li>Sensor / Activity to control loop coordination </li></ul><ul><li>Evolving degree of human involvement </li></ul><ul><li>Cougaar </li></ul><ul><li>Agent societies </li></ul><ul><li>Cougaar environment </li></ul><ul><li>Agent coordination </li></ul><ul><li>Transitioning of control loops human to automation </li></ul>architectural mapping
  12. 12. LMS – ILE Environment Agent Agent Coordination Artifact (CA) Agent Defines roles Agent Role-players Shared state <ul><li>Coordination Artifacts: CAs </li></ul><ul><ul><li>Are first-class entities in MAS </li></ul></ul><ul><ul><li>Define explicit roles for role-players </li></ul></ul><ul><ul><li>Offer shared state between the role-player & the CA </li></ul></ul><ul><ul><li>Coordinate behavior among role-players </li></ul></ul><ul><ul><li>Have distributed implementation </li></ul></ul>
  13. 13. Dynamic Re-Planning and Execution Monitoring <ul><li>Negotiate between agents to iteratively improve plan </li></ul><ul><li>Repetitively recheck assumptions : Constraints, Data, Policies, Metrics </li></ul><ul><li>Updates from data sources as time progresses </li></ul>Data Source Real-Time Data Fusion Self-Assessment Allocation Results Reallocations Task Allocations
  14. 14. Supporting Adaptation in the System Life Cycle Metric Service Management Society Agent Services Coordination Run Environment Configuration Rules Society Configuration Rules Deploy Components Binders/Aspects Plug-ins Configure Event Driven SOA (Service Oriented Architecture) Data Driven Blackboard Knowledge Rep Program Environment Agent Development Phase IDE Application Plug-ins Deploy Rules Spec Tool Run Server Cougaar Middleware Society Monitor
  15. 15. <ul><li>Thank you for your attention </li></ul>