A Next Generation Knowledge Management System Architecture

1
A Next Generation Knowledge Management System
Architecture

• Authors Sameera Abar, Toru Abe, Tetsuo
  Kinoshita
• Source
• Proceedings of the 18th International Conference
  on Advanced Information Networking and
  Application (AINA04)
• ?? ???
• Reporter ???(???)
• Date 2007/12/26

2
Outline

• Introduction
• Knowledge Management
• Architecture of ADIPS-based Knowledge Management
  System
• Concluding Remarks

3
INTRODUCTION

• Different knowledge intensive tasks need
  knowledge from different sources and in different
  presentation formats.
• The unstructured nature of the knowledge
  resources in an organization creates an
  information overload problem.
• This research designs to build an efficient next
  generation active support KM infrastructure that
  automatically streamlines and expedite the
  knowledge related activities in an easy and cost
  effective manner.

4
Knowledge Management(1/2)

• The information that represents the best response
  to a users need is quite often the synthesizing
  of data from a variety of distributed and
  heterogeneous sources.
• These conventional systems are inflexible and
  stifle knowledge creation.
• At present, machines and software can store the
  information, rank it, display it, but cannot
  comprehend or process it.

5
Knowledge Management(2/2)

• The next generation KM systems utilize artificial
  intelligence in the form of intelligent agents as
  well as active knowledge resources represented
  with semantically rich metadata to improve
  inaccuracies of information retrieval, and
  maximize the shareability of knowledge.
• We propose a generic knowledge management system
  architecture based on ADIPS (Agent-based
  Distributed Information Processing System)
  framework.

6
Architecture of ADIPS-based KM System(1/8)
7
Architecture of ADIPS-based KM System(2/8)Interfa
ce Layer

• At the interface layer, the KM system users
  interact with the system to create, explicate,
  use, retrieve, and share knowledge.
• The Personal Assistant (PA) agent represents the
  interests of the user and provides the interface
  between user and the system.
• It gradually learns how to better assist the user
  by observing and imitating the user,
  understanding users interests and needs, and
  building up users profiles.

8
Architecture of ADIPS-based KM System(3/8)Intelli
gence Layer

• Agents can check of the dynamic conditions of the
  knowledge management environment, reason to
  interpret those perceptions, solve problems,
  determine actions, and finally act accordingly.
• Some agents have an ability to learn from past
  mistakes at an explicit level which is something
  very much in line what a KM system is intended to
  help with.

9
Architecture of ADIPS-based KM System(4/8)Structu
red Resource Layer

• The considerable size of the information space
  and the variety of resources residing in it, make
  network information access a daunting task.
  Therefore, knowledge should be organized by an
  appropriate taxonomy for the ease of its
  retrieval.
• By enhancing the existing information sources
  with meta-data, the agents are able to recognize
  and understand what information is and what it is
  about.

10
Architecture of ADIPS-based KM System(5/8)Workpla
ce KM-Agents

• Personal Assistant Agent (PA)
• collaboration with other workplace agents and
  reasoning over the suggested information.
• responding to users query proactively.
• actively updating its knowledge based on the
  information
• fetched and users response.

11
Architecture of ADIPS-based KM System(6/8)Workpla
ce KM-Agents

• Task Management Agent (TMA)
• behaving like a manager agent to handle the
  organization of all other agents which take part
  in some specific KM task.
• administrating and controlling the collaboration
• among users and agents during the
  execution of a task.
• Information Processing Agent (IPA)
• retrieving and merging information from
  heterogeneous distributed information sources.
• filtering irrelevant content under information
  overflow
• condition.

12
Architecture of ADIPS-based KM System(7/8)Workpla
ce KM-Agents

• Resource Agents (RAs)
• protocols availability through which knowledge
  resources accepts queries.
• extracting relevant information for a given
  request.
• managing the status of whole knowledge
  repository.
• actively proposing resources to other agents
  based on their knowledge of other agents needs.

13
Architecture of ADIPS-based KM System(8/8)
14
Concluding Remarks

• The work presented in this paper intends to
  design an adaptive knowledge management
  environment based on the multi-agent technology,
  for providing the users with timely access to the
  just-in-time and context-dependent knowledge with
  an effective approach to managing distributed
  information systems.
• The concept is to specialize agent interactions
  for autonomously and flexibly managing the
  operational knowledge of network devices, thereby
  reducing the workloads of a networks
  administrator remarkably.