Building The Knowledge Grid Building The Knowledge Grid

Document Sample
Building The Knowledge Grid Building The Knowledge Grid Powered By Docstoc
					 Building The
Knowledge Grid
    David Forslund
    David Forslund
   Cognition Group
   Cognition Group
    HealthGrid IRT
    HealthGrid IRT
    March 1-2, 2006
    March 1-2, 2006

  • What is the Knowledge Grid?
  • What is the value of the Knowledge Grid?
  • Sample architecture of the Knowledge Grid
  • Examples:
      - Biosurveillance (Federation)
      - Healthcare Service Specification Project (Services)
      - caBIG/LexGrid/LexBIG (Ontologies)
  • Examples from the EU
  • Challenges for the Knowledge Grid

CognitionGroup, 2006
          What is the Knowledge Grid?

  • Taking the Grid approach beyond data and information to
  • Autonomous distributed computing environment with semantic
    “understanding” (self-organization, adaptation and
  • Ability to link heterogeneous resources semantically to solve
    relevant problems (complex workflow)
  • Related to the Semantic Web effort (documents -> services)
  • An abstract service-based Grid architecture that does not limit
    the user in developing and using service-based knowledge
    discovery applications (Congiusta, Pugliese, Talia, Trunfio

CognitionGroup, 2006
         Knowledge Grid supports Knowledge

                               Grid Services


                       Data mining             Data Archives

CognitionGroup, 2006
               What is the value of the
                 Knowledge Grid?

  • Access to specific information on a patient or disease
    without human intervention
      - Service discovery (ontology)
      - Semantic matching (problem solving ontology)
      - Distributed knowledge workflow
  • e.g., Virtual Patient Record capability (1996-7,
  • Social network for problem solving

CognitionGroup, 2006
       Features of the Knowledge Grid
                       (from Cheung and Liu, 2005)
                       (from Cheung and Liu, 2005)

  • Resource/Service/Domain/Problem Solving/Sociology and Psychology
  • Resource Brokering Agent (RBA)
  • Service Brokering Agent (SBA)
  • Association Discovery Agents (ADA)
  • Data Transformation Agent (DTA)
  • Knowledge Discovery Service Agent (KDA)
  • Service Flow Planning Agent (SFPA)
  • Personalization and Context-aware Agent (PCA)
  • Service Flow Execution Agent (SFEA)
  • Appropriate Security and Privacy constraints

CognitionGroup, 2006
An Architecture of the Knowledge Grid
                       (Cannataro and Talia)
                       (Cannataro and Talia)

CognitionGroup, 2006
       Another view of the Knowledge Grid
                       (von Laszewski, et al.)
                        (von Laszewski, et al.)

CognitionGroup, 2006
                 Requirements for the
                   Knowledge Grid

  • Better agreements (standards?) on creating, negotiating and
    translating ontologies.
      - cf., complexity of UMLS, various ontology representations
  • Better standards to move from workflow to knowledge flow
      - BPEL, XPDL, …
  • Robust mechanisms for negotiating security and privacy policy
  • Adaptable technology base (XML isn’t forever, REST, …)
  • Proper division and exposure of services

CognitionGroup, 2006
            Federated BioSurveillance

  • B-SAFER: Example of using service components to
    build a system for medical surveillance
  • Leverage several standards (OMG & HL7) for
      - EHR component adapted to surveillance
  • Architecture enables federation of surveillance so that
    local systems can be autonomous but provide data to
    a higher level
  • Facilitates HIPAA compliance
CognitionGroup, 2006
        What is the Healthcare Service
           Specification Project?

      • An effort to create common “service interface specifications”
        tractable within Health IT
      • A joint standards development project involving Health Level
        7 (HL7) and the Object Management Group (OMG)
      • Its objectives are:
          - To create useful, usable healthcare standards that address
            functions, semantics and technologies
          - To complement existing work and leverage existing standards
          - To focus on practical needs and not perfection
          - To capitalize on industry talent through open community

CognitionGroup, 2006
      Why “common services” and not
            just “messages”?*
   • A common practice in healthcare, just not yet in healthcare IT
   • Many key products use them but do not expose interfaces
   • Ensures functional consistency across applications
   • Accepted industry best practice
   • Furthers authoritative sources of data
   • Minimizes duplication across applications, reuse
   • Messages can be either payloads in or infrastructure beneath
   • Service-oriented architecture is just automation of common
                       *slide adapted from a Veterans Health Administration Presentation, used with permission
CognitionGroup, 2006
              What is being specified in

      • Standards are being developed for:
      • Entity Identification
          - to manage and maintain identities within and across domains, localities, or

      • Record Location & Retrieval
          - to discover, retrieve, and update records in distributed environments

      • Decision Support Services
          - to support evaluation processes such as clinical decision support

      • Terminology Service
          - to retrieve, maintain, and navigate [clinical] terminologies and ontologies)

CognitionGroup, 2006
           LexGrid Components provide
                 flexible sevices
                                 Browse and
                       Import                            Query

            Export                                S
                           DataStore              e               Web
                                                  r              Clients
                                       Embed                        Java
                                                  i                 .NET
           CSV                                    c
  OWL                                                                  ...
     RDF         ...                              s
CognitionGroup, 2006
           The Cancer Bioinformatics Grid

  A voluntary network or grid connecting individuals and
    institutions to enable the sharing of data and tools, creating a
    World Wide Web of cancer research. The goal is to speed the
    delivery of innovative approaches for the prevention and
    treatment of cancer.

CognitionGroup, 2006
                       Project Goals

  • Build a vocabulary server
       - Accessible through an application programming
         interface (API)
       - Based on Standards
       - Using Commodity Technologies
       - With import/export functionality

CognitionGroup, 2006
                       LexBIG Vision

                                      Online Replica

                                      Server           NCI


                                                                Local Replica

                       caBIG                                    I

                        Grid                                    o
                                                                              NCI Meta-

                                                                e               Other
                                                                r     CTS
                                                                     Server         NCI


                                  NCI Meta-         m
                                  Thesaurus         p
                       LexGrid    Vocabulary
                        CTS                         i
                       Server          NCI          t

CognitionGroup, 2006
           Uses of the Knowledge Grid
              Identified in the EU

  • Integrating Health Information
  • Improved Medical Knowledge
  • Functional systems

CognitionGroup, 2006
        Integrating Health Information
                                Public Health

                                  Medical Imaging


CognitionGroup, 2006
          Improved Medical Knowledge

  • Patient specific computational models of anatomy and
    physiology for disease modeling and simulation from
    molecular level to organ level e.g. new approaches to
    drug discoveries
  • Methods to map and seamlessly link clinical and genetic
    information resources e.g. using grid technologies,
    electronic health records including genetic information
  • ICT methods for “in vivo” visualisation of biological
    processes on molecular level (ICT in support of
    molecular imaging)

CognitionGroup, 2006
              Functional systems (EU)

  •     Interoperability of eHealth systems – realistic
        approaches to this concept with clinical applicability.
        Special emphasis on semantic interoperability and the
        further R&D needed in the area of biomedical
  •     In silico model of human being (virtual or
        physiological human) from eCell to eOrgan.

CognitionGroup, 2006
                 Example: MammoGrid
               Why a Grid infrastructure ?
                         (Norager, EC)
                         (Norager, EC)

 • Large federated databases
    - large data (size of images)
    - needs enough data for statistics
 • Ontologies and metadata
    - Heterogeneous image formation parameters and features
    - Needs “standard” for comparison
    - Clinical information
    - Geographically distributed data
 • Effective data mining of a rapidly growing database
 • Computing expensive simulations
 • Allow for complex queries involving executables

CognitionGroup, 2006
                  MammoGrid Challenges
                                  (Norager, EC)
                                  (Norager, EC)

  • Data resides in hospitals
     - Firewall protected
  • Legal restrictions on access to data
     - Clinicians, researchers, developers, Govt, …
  •   Secure file transfer - Patient privacy & security
  •   Combining several databases
  •   Medical image analysis clients are not Grid experts!
  •   Services must be system-resident, invisible, generic.

CognitionGroup, 2006
   Impact of Grid technology in Health
                           (Norager, EC)
                           (Norager, EC)

  • Define collaborations.
    There is a need for both technology developers and users
    in order to move from “research to applications”.
  • Define resources to be shared and access rights - address
    privacy and security issues.
  • Awareness of other Grid’s - no reinventing the wheel,
  • Creation of a HealthGrid community will ensure impact
    on standards, enable showcases etc.

CognitionGroup, 2006
       To be successful KG must be                           (Talia)

  • Easy to program - hiding architecture issues and details,
  • Adaptive - exploiting dynamically available resources,
  • Human-centric - offering end-user oriented services,
  • Secure - providing secure authentication mechanisms,
  • Reliable - offering fault-tolerance and high availability,
  • Scalable - improving performance as problem size increases,
  • Pervasive - giving users the possibility for ubiquitous access,
  • Knowledge-based - extracting and managing knowledge
    together with data and information.
CognitionGroup, 2006
    HealthGrid challenges identified in
                 the EU
  • Addressing research and healthcare
     - Semantic integration -building on the results
       related to interoperability of clinical data and
       research data
     - Very heterogeneous user community -health
       professionals, researchers, authorities, patients
     - New combination of data –ex. biomedical
  • Distributed sources of data, some with strong
    privacy rules
      - Algorithms for search, data mining and knowledge

CognitionGroup, 2006
       HealthGrid challenges identified
               in the EU cont.

  • Privacy and Security
     - Adapting existing solutions to the Grid
     - High reliability of system must be proven
  • User friendliness and acceptance
     - Need to provide fast and easy to use tool at the point of
     - Organizational and cultural issues related to new ways of
       working and virtual collaborations
  • From research technology to a solution in Healthcare
      - Political awareness and understanding
      - Business models
      - Legal and ethical issues

CognitionGroup, 2006

  • Grid provides robust infrastructure for knowledge
  • Semantic integration remains a challenge particularly
    for reasoning required for knowledge
  • Security policy is a human (legal, organizational and
    cultural) problem more than technology
  • Usability within a given application area remains an
  • International integration challenges

CognitionGroup, 2006
  •   A Framework for Building Scientific Knowledge Grids Applied to Thermochemical Tables
       -   Gregor von Laszewski, Branko Ruscic, Kaizar Amin, Patrick Wagstrom, Sriram Krishnan, and Sandeep Nijsure
  •   Enabling Knowledge Discovery Services on Grids
       -   Domenico Talia
  •   On Knowledge Grid and Grid Intelligence: A Survey
       -   William K. Cheung and Jiming Liu
  •   KNOWLEDGE GRID: An Architecture for Distributed Knowledge Discovery
       -   Mario Cannataro and Domenico Talia
  •   An International Collaboratory Based on Virtual Patient Records
       -   Kilman and Forslund
  •   The Vision of a HealthGrid
       -   Sofie Norager
  •   Setting Standards for Improved Syndromic Surveillance
       -   Forslund, Joyce, Burr, Picard, Wokoun, Umland, Brillman, Froman, Koster
  •   LexGrid and caBIG (aka LexBIG)
       -   Harold Solbrig
  •   Integrated Biomedical Information for Better Health
       -   IST Conference 2005 (EU)
  •   HealthGrid Healthcare & Research on the Grid
       -   GérardComyn

CognitionGroup, 2006

Shared By:
Tags: Knowledge, Grid
Description: Fran Berman earlier proposed the concept of the knowledge grid, knowledge grid is an intelligent interconnection environment that enables users or virtual roles to effectively capture, publish, share and manage knowledge resources, and other services for the users and to provide the required knowledge services, support for knowledge innovation, and work together.