Agent technology Enabling communication among tools and data

Agent technology Enabling communication among tools and data Thomas E. Potok Collaborative Technologies Group Leader Computer Science and Mathematics Division Oak Ridge National Laboratory Mladen Vouk Professor of Computer Science Technical Director of the NC State Center for Advanced Computing and Communications North Carolina State University

Agent Technology Trend TS N GE A Internet Telephone Face to Face Tim Berners-Lee – “The Semantic Web… will open up the knowledge and workings of humankind to meaningful analysis by software agents, providing a new class of tools by which we can live, work and learn together” “The Semantic Web” Scientific American 5/2001

Successful Projects VIPAR Knowledge Discovery }We have extensive expertise in Supply Chain Management Agent System agent development Manufacturing Emulation Agent System }Began working with agent Collaborative Decision Support System technologies in 1980 s Neural Nets for Recovery Boiler Control Neural Nets for Bankruptcy Prediction Neural Nets for Spring-back Prediction Collaborative Design System Neural Nets for Resistance. Spot Welding Neural Nets for Material Mix Optimization Genetic Algorithms for Chemical Synthesis Knowledge-based Systems - Manufacturing Advisors Knowledge-based Systems for Constructability Design and Analysis of Computer Experiments Knowledge-based Computer Systems Calibration 1985 1990 1995 }Over 10 successful projects within the last 5 years }Collaborations with leading agent experts 2000

What are Agents? . . . Software entities that assist people and act on their behalf - IBM. . . Software “robots” Traditional Software Object Proactive detect changes in their environment and react to those in a timely manner by answering to events and initiating actions Behavior Goal-driven have a purpose and act in accordance with that purpose until it is fulfilled Communicative able to interact and communicate with users and other agents State Autonomous Learning have the ability to learn from experiences in their environment can have control over their own actions and be able to work and launch actions independent of the user or other actors

Typical Agent Architecture MULTIAGENT APPLICATION (Agent-User Interface, Conversation Interface) Legacy Databases Legacy Software INFERENCE ENGINE (Decision Support Models) LOCAL MODEL (Domain KB, Ontologies) SOCIAL MODEL (Interaction rules, Conversation Management) LINGUISTIC LAYER (Message Syntax and Semantics) COMMUNICATION PROTOCOL LAYER (Multicast, Directed Communication) LEARNING LAYER INFRASTRUCTURE (e. g. , Java Virtual Machine) OPERATING SYSTEM HARDWARE

Current Agent Funding · NSF: - Division of Information and Intelligent Systems - Division of Electrical and Communications Systems · Computation and Social Systems · Knowledge and Cognitive Systems · Control, Networks, and Computational Intelligence (CNCI) · DARPA: - Information Technology Office · Control of Agent-Based Systems (Co. ABS) · DARPA Agent Mark Up Language (DAML) · Taskable Agent Software Kit (TASK) · Office of Naval Research: - · Information, Electronics & Surveillance · Intelligent Systems Others include: - NASA, Army, Air force, Office of the Secretary of Defense, NIST, …

Simple Agent Example 2 Days $18. 50 Barnes and Nobel 1 Day $21. 75 1 Day $20. 25 Library 1 Day Free e. Bay Dedicated Agents Amazon B. Dalton Agent, find me the book “War and Peace, ” and I need it tomorrow 1 Day $12. 50 Agent Communities Does the agent understand buying books? Form a plan to buy the book Execute the plan

VIPAR Project – US Pacific Command • Analysts Search 20 -30 papers a day • Manually find the articles on internet • Manually organizing the articles Inside China Unorganized Newspaper Articles Analysts Organized Results NEA Russia Today SEA Jakarta Post Oceania

New Intelligent Approach Intelligent Agent Inside China Analysts Economy Russia Today Military Jakarta Post Political Retrieval Agents Vector Perturbation Clustering “Knowledge Bins”

Oak Ridge Mobile Agent Community (ORMAC) Communication Machine A Machine B Agent Host X Agent Host Y Agent Host Z Agent Context Agent 1 Agent 2 Agent 3 · Allows agents to be transmitted and received among machines. · Agents can also interact with systems and agents that are not part of the community. · Agents uses the Foundation for Intelligent Physical Agent (FIPA) compliant agent control language (ACL) messages.

VIPAR Architecture RDF Conversion from Dynamic Vector Ontologies HTML to XML Space Models Information Agents Whiteboard Agent GTP Agent Information Agents Cluster Agent … Information Agents External Agents … External Agents Multi-agent System Built on ORMAC Framework

Clustering: A Simple Example 1. Original and Parsed Documents 2. Terms Dictionary 3. Vector Space Model 4. Document Similarity 5. Euclidean Distance 6. Agglomerative Hierarchical Clustering 7. Phylips tree representation 1 Divisive Agglomerati ve D 1 2 3 4 8 6 7 D 2 D 3

VIPAR Knowledge Management Same result, easier to “see” 3 2 v Wen Ho Lee 1. 2. 5 v Clinton Concerned Over Lee Case - Reno On Defensive Asian-Americans Demand White House Action on Lee 5. Wen Ho Lee Clinton Calls For Review Of Lee Secrets Case 4. 1 Wen Ho Lee Spends First Day Savoring Home Delights 3. 4 Lee Case Points up Scientists' Attitude on Security India and Pakistan 6. 7. US China Trade India/Pakistan v India and Pakistan: Troubled relations Troops die in Kashmir clashes IAEA Meeting 8. IAEA Meeting 10 v 7 IAEA Supports Putin Nuclear Power Initiative 9. 6 China Rejects Moves to Tighten Regulation of Nuclear Materials U. S. China Trade 10. 8 9 U. S. China Trade Vote Milestone on Rocky Road

Endorsements · US Pacific Command’s CINC Admiral Blair - “A tremendously successful project” - “Software agents … lead to substantially improved analytical products. ” · US Pacific Command’s Science and Technology Advisor - “hit a grand slam home run!” - “first time we've seen information discovery and knowledge management software working here at HQ USCINCPAC operationally. ”

How agents can work for scientific data

Astrophysics simulation of temperature and tangential velocities Visualization created by Ross Toedte

Hypothetical Scenario Probe Astrophysics Simulation Astrophysics Data Agent Storage Agent Parallel IO Agent Stripe Analysis Data Reduction Agent Binning Sampling Clustering Agent Dynamic Clustering MPI-IO Full Data HPSS Reduced Data Clustered Data

Example SDM Agent Community Scientist: What if we reduce the data? A 1 Agent whiteboard A 2 A 3 A 4 A 5 A 1: REQUEST All: Dimensionality Reduction FORMAT: HDF 5 XML TYPE: Climate A 2: REPLY A 1: Dimensionality Reduction FORMAT: HDF 5 TYPE: General A 3: REPLY A 1: Dimensionality Reduction FORMAT: ASCII TYPE: Physics A 1: REQUEST All: Conversion FORMAT: HDF 5 XML to HDF 5 TYPE: any A 4: REPLY A 1: Conversion FORMAT: HDF 5 XML to HDF 5 TYPE: Nuclear A 5: REPLY A 1: Conversion FORMAT: HDF 5 XML to HDF 4 XML TYPE: any A 1: PERFORM A 4: Conversion FORMAT: HDF 5 XML to HDF 5 DATA: D 1 A 4: RESULT A 1: Conversion FORMAT: HDF 5 XML to HDF 5 Data: R 1 A 1: PERFORM A 2: Dimensionality Reduction FORMAT: HDF 5 Data: R 1 A 2: RESULTS A 1: Dimensionality Reduction FORMAT: HDF 5 Data: R 2

Year 1 Deliverables · Generic scientific data ontology - Current framework supports a standard FIPA message ACL - Need for agents to be able understand interact with scientific data · Data centric? · Tool centric? · Analysis centric?

Year 1 Deliverables continued · Generic data and model agents for data mining and discovery of access patterns - Define general case agents that can be adapted to work with · Specific data types · Specific tools · Establish network-based software configuration, management and distribution site

Summary · Intelligent agents are a significant future technology direction · We have a significant progression of deployed agent systems and research to our credit · We have ringing endorsements regarding the success of our systems and research · We look forward to the next steps…

How to use agents? · We do not what to redo what you have already done!!! · We want to supply the data communication protocols so that scientists can use your tools more effectively · Java based, FIPA compliant