Critical Reading and Analysis Example Paper: Brain Meets Brawn: Why Grid and Agents Need Each Other - PowerPoint PPT Presentation


Title: Critical Reading and Analysis Example Paper: Brain Meets Brawn: Why Grid and Agents Need Each Other


1
Critical Reading and AnalysisExample Paper
Brain Meets Brawn Why Grid and Agents Need Each
OtherI. Foster, N. Jennings, C. KesselmanProc.
Autonomous Agents and Multi-Agent Systems
Conference 2004, New York.
  • Module Co-ordinators
  • Shamima Paurobally (paurobs_at_wmin.ac.uk)
  • Radmila Juric (R.Juric_at_westminster.ac.uk)

PPPP slides for Monday 22nd October 2007 (wk1
slides 2)
2
Reason for choosing this paper
  • I. Foster, N. Jennings, C. Kesselman. Brain Meets
    Brawn Why Grid and Agents Need Each Other.
    Proc. Autonomous Agents and Multi-Agent Systems
    Conference 2004, New York.
  • The paper gives a review of the state of the art
    in Grid and Agents
  • Contribution shown as a requirement for
    integrating Grid and agent
  • Solution presented
  • One of the first papers to propose a bridge
    between 2 research areas

3
Objectives when Studying this paper
  • Point out what literature reviews looks like
  • Example of good structure
  • Analyse each sections aims
  • Guidelines when you are writing your report
  • Review this paper as an example of what you are
    expected to do in task 1 (group literature
    review)
  • Our comments are in green

4
Foreword
  • What are agents
  • Distributed Artificial Intelligence
  • Autonomous, reactive, intelligent, proactive,
    social, learning, collaborative,
  • What are Grid/web services
  • Parallel large scale resource sharing, cluster
    computing
  • Remote invocation of a service through a
    published interface

5
Abstract
  • The Grid and agent communities both develop
    concepts and mechanisms for open distributed
    systems, albeit from different perspectives.
  • The Grid community has historically focused on
    brawn infrastructure, tools, and applications
    for reliable and secure resource sharing within
    dynamic and geographically distributed virtual
    organizations. In contrast, the agents community
    has focused on brain autonomous problem
    solvers that can act flexibly in uncertain and
    dynamic environments.
  • Yet as the scale and ambition of both Grid and
    agent deployments increase, we see a convergence
    of interests, with agent systems requiring robust
    infrastructure and Grid systems requiring
    autonomous, flexible behaviors.
  • Motivated by this convergence of interests, we
    review the current state of the art in both
    areas, review the challenges that concern the two
    communities, and propose research and technology
    development activities that can allow for
    mutually supportive efforts.

Note how the whole abstract flows smoothly
6
What can we infer from the abstract
  • One sentence introduction
  • A couple of sentences on the domain(s) Grid and
    Agent
  • Problem statement requirement of each domain
  • The contribution of this paper review of state
    of the art, and proposed solution

7
Analysis of the Introduction
  • Introduction re-enforces the abstract, but in
    more detail
  • Can repeat some sentences from the abstract
  • Present the background context
  • Open distributed systems, covering Grid and
    agents
  • Virtual organisations as the common factor
  • Identify the problem in the domains
  • Scale and application in future generation
  • Grid rigid and inflexible interoperation and
    interactions
  • Agent small systems, not robust and secure
  • Need for each other. Grid seeks to be flexible
    and agile, Agent seeks to be reliable and
    scalable
  • Contribution of the paper
  • Examine work in the two domains and motivation
    for this is inform each community and cross
    fertilisation
  • Paper is a first step towards that goal (here
    stating how it is advancing the state of the art)
  • Structure of paper (in terms of sections)
  • Mention scope to counter criticisms (optional)

8
State of Art of Grids - Standards
  • Overall objectives of Grids for resource sharing
    in VOs
  • Standardisation of protocols and interfaces
  • Technologies (Standards)
  • Early ad hoc frameworks -gt GT (Globus Toolkit)
    standards -gt web services standards
  • OGSA and WSDL interfaces for defined, discovered
    and invoked WS
  • WSRF, WS-Agreement
  • GT and VOs, service oriented architectures
  • BUT control mechanisms for dealing with failures
    and adaptiveness to changing environmental
    conditions and application concerns are missing!

9
State of Art of Grids - Applications
  • Example of review of Grid Applications
  • Early applications scientific computing
  • Large-scale distributed computing, data grids
  • Now uptake in industry for distributed systems
  • GT in VOs integrating resources (statistics
    given) and future goal of 1000s of sites
  • NEESgrid for earthquake simulation
  • 3 sites, 50 remote participants
  • Grid3 28 sites, 3000 processors
  • AccessGrid video conferencing
  • Butterfly.net GT-based for multiplayer online
    games
  • GlobeXplorer GT for satellite image data
  • BUT experiences reveal issues that must be
    addressed if Grids are to be scaled to larger
    communities, more diverse resources, and more
    complex applications

10
Agent-Based Computing
  • Definition and characteristics of an agent (with
    references for more details)
  • Multi-agent Systems for most problem solving
    (LINK to VO)
  • Interaction to achieve goals (social welfare or
    individual)
  • Service discovery and invocation (LINK to Grid)
  • In addition, sophisticated social interactions
  • Cooperate, coordinate, negotiate (which Grid does
    not have)
  • Relationship between agents VOS and teams (LINK
    to Grid)
  • What differs agents from Grids
  • Sophisticated agent interactions
  • Flexible and adaptiveness to environment, partial
    knowledge and control
  • BUT with autonomy and flexibility, it is
    difficult to ensure that desirable global
    behaviours emerge and need to impose greater order

11
Agent Technologies and Applications
  • Authors say, more existing work on agent
    theories, models and algorithms
  • Individual agent, planning in dynamic
    environments
  • Balance between being too responsive and too
    committed
  • Agent architectures
  • JACK, JADE, Cougaar, Zeus
  • Gaia, Tropos, AUM
  • FIPA, KQML
  • BUT, as in Grids, increasing reliance on Web
    services and semantic web for providing
    computational infrastructure and acceptance of
    trust as a central issues in interaction
  • Applications Agent technology deployed over last
    decade, increase in agent applications, deployed
    applications in domains such as manufacturing,
    electronic commerce, process control,

12
Our critical Analysis of Section 3
  • Good overview of what an agent is and links
    between agent VOs and Grids
  • Technologies are sparsely presented
  • They say that most problems require multiple
    agents, but then focus on a single agent in
    technologies
  • Claim that more on theory and models fine, but
    there are still agent technologies that they
    briefly list in one paragraph, without any
    explanation or review of these technologies
  • Some of the technologies are not widely adopted
  • Trust is not such a central issue in agents as
    they claim (about 5)!
  • SHOULD have concentrated on the popular
    technology in more detail
  • Applications subsection is very thin!
  • Too general, no examples of early or existing
    applications
  • No references to applications in the mentioned
    domains to validate their claims

13
Section 4 Brains and Brawns
  • Defines the problem this paper brings to light in
    the research community
  • VOs is common thread between Grids and agents
  • Focused on different aspect.
  • Grids Community standards and policy enforcement
  • Problem in Grids use of mechanisms to create
    large-scale systems with stable collective
    behaviour is less mature
  • For example, commonly used Grid tools provide
    uniform mechanisms for accessing data on
    different storage systems, but not for the
    semantic integration of that data for accessing
    service and resource state, but not for
    anticipating, detecting, and diagnosing problems
    implied by changes to that state and for
    securely authenticating users and services, but
    not for inferring whether or not specific users
    or services can be trusted to perform specific
    actions. To this extent, Grids are all brawn and
    no brain.

14
Section 4 Brains and Brawns
  • Agents also focus on creating community (link to
    Grid)
  • Flexible decision making, rich social
    interactions
  • Problem However in building all this flexibility
    and sophistication, scant attention has been paid
    to how these tasks should be performed in
    realistic distributed environments. For example,
    agent frameworks provide sophisticated internal
    reasoning capabilities, but offer no support for
    secure interaction or service discovery
    cooperation algorithms produce socially optimal
    outcomes, but assume the agents have complete
    knowledge of all outcomes that any potential
    grouping can produce and negotiation algorithms
    achieve optimal outcomes for the participating
    agents, but assume that all parties in the system
    are known at the outset of the negotiation and
    will not change during the systems operation.
    Thus, one may say that agents are all brain and
    no brawn.

15
Re-stating the Problem for Both
  • From these problems, re-enforcing the points
    about Grids and Agents needing each other
  • Stating that the simple approach of layering is
    not enough
  • Need fine-grain intertwining for true benefits
    (stating that that is not a problem with a simple
    solution)
  • Early examples of solutions
  • Agent-based resource selection through
    re-inforcement learning
  • Reference to a paper at the same conference that
    this paper is published.
  • Second example is the use of automated
    negotiation techniques to allocated Grid
    resources
  • But the work referenced here is not about agents.
  • Not much about integration, just mention a couple
    of examples, and that this will bring challenges
    but what challenges?

16
Section 5 Robust Agile Service Oriented
Architectures
  • Draw the two we now draw the two parallel lines
    of research together to highlight their
    commonalities and complementarities.
  • Did not section 4 do that already?
  • Service as unifying concept (Autonomous services)
  • Wasnt VO the unifying concept?
  • Replicating databases and Distributed negotiation
    protocols might be used to establish the query
    throughput achievable on individual copies, such
    that community throughput is optimized.
  • Replicating databases is not so straightforward
    as the authors think. They have disregarded the
    large body of existing work on databases,
    consistency, and synchronisations.
  • The sentence on distributed negotiation protocols
    for querying databases seems unrealistic
  • Distributed planning and scheduling already
    exists in Grids, as is disregarded by the paper
  • The example of the database service is not
    convincing (blue sky thinking)

17
Section 5.2 Rich Service Models
  • Contribution of Grid is a robust lifetime and
    naming model for dynamic services
  • Service failure and scalability to benefit agent
  • But the paper mentioned before that one of the
    weakness of Grids is inability to deal with
    failure
  • Statefulness of Grids (persistency of states
    over lifetimes)
  • In contrast, agent systems address semantics but
    do not provide a consistent state model.
  • This statement should be explained, may be with
    an example and justified
  • No real solution in this section of how to
    combine the semantics of agent with state of
    services
  • Different types of semantics in agent and in Grids

18
Negotiation and Service Contracts
  • Negotiation of level of service provision
  • To establish service contracts
  • Well explained section with example of
    negotiation issues e.g. load, identity, etc.
  • What is a Byzantize failure model no reference
    here
  • Grids has a promising approach of having shared
    policy statement to represent an agreement
  • Too brief mention of strategy

19
Virtual Organisation Management
  • VOs in both Grids and Agents
  • VO creation and operation
  • Involves negotiation
  • Much work on VO in agents
  • However no reference, examples given
  • Given importance of VOs given earlier in paper,
    this section is too brief

20
Authentication, Trust and Policy
  • First paragraph from a Grid point of view,
    association of identity and community based
    authorisation
  • Second and third paragraphs Trust, negotiate
    trust and policy
  • Why is trust mentioned here?
  • Claims about expecting Grids to use agent trust
    and reputation is not justified
  • Not coherent with first paragraph
  • Trust is given too much importance with respect
    to existing work in agent community and with
    respect to the needs of the Grid
  • In fact after 3 years, scarcely any work on Grid
    and trust exists i.e. trust is not currently
    important for Grid

21
Ten Research Problems
  • Some kind of proposal for a solution towards
    Agents and Grid integration for future work
  • Remember this is a position paper not an
    implementation/theoretical paper.
  • After 3 years, which of these research problems
    are being tackled?
  • Integrated service architecture
  • Interesting, but does not seem to work after 3
    years. Few agent grid service architectures
    proposed and even less adopted
  • Trust negotiation and management
  • Again why is trust given so much importance?
  • Few Grids are concerned with trust negotiation
  • System management and troubleshooting
  • Where is the contribution of agents mentioned
    here?

22
Ten Research Problems
  • Negotiation
  • Mentioned too briefly?
  • Service composition, VO formation and management,
    System predictability
  • Seems promising but again how can agents help
    here/integrate here? No mention of agents
  • Human-computer collaboration
  • Where is the agent and where is the Grid?
  • Evaluation
  • Promising and well-explained how agent Grid can
    integrate to solve this problem
  • Semantic integration
  • Where is the Grid and agent here?

23
Critical Overview
  • First part of paper well described
  • Latter part of paper seems to be written by
    different authors and not coherent with each
    other
  • Ten research problems
  • Most written by Grid person, a couple on trust by
    agent person
  • No coherence
  • No mention of how agent Grid integration can help
    in most of these research problems
  • Conclusions is the research challenge problems

24
Questions
View by Category
About This Presentation
Title:

Critical Reading and Analysis Example Paper: Brain Meets Brawn: Why Grid and Agents Need Each Other

Description:

Point out what literature reviews looks like. Example of good structure ... AccessGrid video conferencing. Butterfly.net: GT-based for multiplayer online games ... – PowerPoint PPT presentation

Number of Views:127
Avg rating:3.0/5.0
Slides: 25
Provided by: tasosp
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Critical Reading and Analysis Example Paper: Brain Meets Brawn: Why Grid and Agents Need Each Other


1
Critical Reading and AnalysisExample Paper
Brain Meets Brawn Why Grid and Agents Need Each
OtherI. Foster, N. Jennings, C. KesselmanProc.
Autonomous Agents and Multi-Agent Systems
Conference 2004, New York.
  • Module Co-ordinators
  • Shamima Paurobally (paurobs_at_wmin.ac.uk)
  • Radmila Juric (R.Juric_at_westminster.ac.uk)

PPPP slides for Monday 22nd October 2007 (wk1
slides 2)
2
Reason for choosing this paper
  • I. Foster, N. Jennings, C. Kesselman. Brain Meets
    Brawn Why Grid and Agents Need Each Other.
    Proc. Autonomous Agents and Multi-Agent Systems
    Conference 2004, New York.
  • The paper gives a review of the state of the art
    in Grid and Agents
  • Contribution shown as a requirement for
    integrating Grid and agent
  • Solution presented
  • One of the first papers to propose a bridge
    between 2 research areas

3
Objectives when Studying this paper
  • Point out what literature reviews looks like
  • Example of good structure
  • Analyse each sections aims
  • Guidelines when you are writing your report
  • Review this paper as an example of what you are
    expected to do in task 1 (group literature
    review)
  • Our comments are in green

4
Foreword
  • What are agents
  • Distributed Artificial Intelligence
  • Autonomous, reactive, intelligent, proactive,
    social, learning, collaborative,
  • What are Grid/web services
  • Parallel large scale resource sharing, cluster
    computing
  • Remote invocation of a service through a
    published interface

5
Abstract
  • The Grid and agent communities both develop
    concepts and mechanisms for open distributed
    systems, albeit from different perspectives.
  • The Grid community has historically focused on
    brawn infrastructure, tools, and applications
    for reliable and secure resource sharing within
    dynamic and geographically distributed virtual
    organizations. In contrast, the agents community
    has focused on brain autonomous problem
    solvers that can act flexibly in uncertain and
    dynamic environments.
  • Yet as the scale and ambition of both Grid and
    agent deployments increase, we see a convergence
    of interests, with agent systems requiring robust
    infrastructure and Grid systems requiring
    autonomous, flexible behaviors.
  • Motivated by this convergence of interests, we
    review the current state of the art in both
    areas, review the challenges that concern the two
    communities, and propose research and technology
    development activities that can allow for
    mutually supportive efforts.

Note how the whole abstract flows smoothly
6
What can we infer from the abstract
  • One sentence introduction
  • A couple of sentences on the domain(s) Grid and
    Agent
  • Problem statement requirement of each domain
  • The contribution of this paper review of state
    of the art, and proposed solution

7
Analysis of the Introduction
  • Introduction re-enforces the abstract, but in
    more detail
  • Can repeat some sentences from the abstract
  • Present the background context
  • Open distributed systems, covering Grid and
    agents
  • Virtual organisations as the common factor
  • Identify the problem in the domains
  • Scale and application in future generation
  • Grid rigid and inflexible interoperation and
    interactions
  • Agent small systems, not robust and secure
  • Need for each other. Grid seeks to be flexible
    and agile, Agent seeks to be reliable and
    scalable
  • Contribution of the paper
  • Examine work in the two domains and motivation
    for this is inform each community and cross
    fertilisation
  • Paper is a first step towards that goal (here
    stating how it is advancing the state of the art)
  • Structure of paper (in terms of sections)
  • Mention scope to counter criticisms (optional)

8
State of Art of Grids - Standards
  • Overall objectives of Grids for resource sharing
    in VOs
  • Standardisation of protocols and interfaces
  • Technologies (Standards)
  • Early ad hoc frameworks -gt GT (Globus Toolkit)
    standards -gt web services standards
  • OGSA and WSDL interfaces for defined, discovered
    and invoked WS
  • WSRF, WS-Agreement
  • GT and VOs, service oriented architectures
  • BUT control mechanisms for dealing with failures
    and adaptiveness to changing environmental
    conditions and application concerns are missing!

9
State of Art of Grids - Applications
  • Example of review of Grid Applications
  • Early applications scientific computing
  • Large-scale distributed computing, data grids
  • Now uptake in industry for distributed systems
  • GT in VOs integrating resources (statistics
    given) and future goal of 1000s of sites
  • NEESgrid for earthquake simulation
  • 3 sites, 50 remote participants
  • Grid3 28 sites, 3000 processors
  • AccessGrid video conferencing
  • Butterfly.net GT-based for multiplayer online
    games
  • GlobeXplorer GT for satellite image data
  • BUT experiences reveal issues that must be
    addressed if Grids are to be scaled to larger
    communities, more diverse resources, and more
    complex applications

10
Agent-Based Computing
  • Definition and characteristics of an agent (with
    references for more details)
  • Multi-agent Systems for most problem solving
    (LINK to VO)
  • Interaction to achieve goals (social welfare or
    individual)
  • Service discovery and invocation (LINK to Grid)
  • In addition, sophisticated social interactions
  • Cooperate, coordinate, negotiate (which Grid does
    not have)
  • Relationship between agents VOS and teams (LINK
    to Grid)
  • What differs agents from Grids
  • Sophisticated agent interactions
  • Flexible and adaptiveness to environment, partial
    knowledge and control
  • BUT with autonomy and flexibility, it is
    difficult to ensure that desirable global
    behaviours emerge and need to impose greater order

11
Agent Technologies and Applications
  • Authors say, more existing work on agent
    theories, models and algorithms
  • Individual agent, planning in dynamic
    environments
  • Balance between being too responsive and too
    committed
  • Agent architectures
  • JACK, JADE, Cougaar, Zeus
  • Gaia, Tropos, AUM
  • FIPA, KQML
  • BUT, as in Grids, increasing reliance on Web
    services and semantic web for providing
    computational infrastructure and acceptance of
    trust as a central issues in interaction
  • Applications Agent technology deployed over last
    decade, increase in agent applications, deployed
    applications in domains such as manufacturing,
    electronic commerce, process control,

12
Our critical Analysis of Section 3
  • Good overview of what an agent is and links
    between agent VOs and Grids
  • Technologies are sparsely presented
  • They say that most problems require multiple
    agents, but then focus on a single agent in
    technologies
  • Claim that more on theory and models fine, but
    there are still agent technologies that they
    briefly list in one paragraph, without any
    explanation or review of these technologies
  • Some of the technologies are not widely adopted
  • Trust is not such a central issue in agents as
    they claim (about 5)!
  • SHOULD have concentrated on the popular
    technology in more detail
  • Applications subsection is very thin!
  • Too general, no examples of early or existing
    applications
  • No references to applications in the mentioned
    domains to validate their claims

13
Section 4 Brains and Brawns
  • Defines the problem this paper brings to light in
    the research community
  • VOs is common thread between Grids and agents
  • Focused on different aspect.
  • Grids Community standards and policy enforcement
  • Problem in Grids use of mechanisms to create
    large-scale systems with stable collective
    behaviour is less mature
  • For example, commonly used Grid tools provide
    uniform mechanisms for accessing data on
    different storage systems, but not for the
    semantic integration of that data for accessing
    service and resource state, but not for
    anticipating, detecting, and diagnosing problems
    implied by changes to that state and for
    securely authenticating users and services, but
    not for inferring whether or not specific users
    or services can be trusted to perform specific
    actions. To this extent, Grids are all brawn and
    no brain.

14
Section 4 Brains and Brawns
  • Agents also focus on creating community (link to
    Grid)
  • Flexible decision making, rich social
    interactions
  • Problem However in building all this flexibility
    and sophistication, scant attention has been paid
    to how these tasks should be performed in
    realistic distributed environments. For example,
    agent frameworks provide sophisticated internal
    reasoning capabilities, but offer no support for
    secure interaction or service discovery
    cooperation algorithms produce socially optimal
    outcomes, but assume the agents have complete
    knowledge of all outcomes that any potential
    grouping can produce and negotiation algorithms
    achieve optimal outcomes for the participating
    agents, but assume that all parties in the system
    are known at the outset of the negotiation and
    will not change during the systems operation.
    Thus, one may say that agents are all brain and
    no brawn.

15
Re-stating the Problem for Both
  • From these problems, re-enforcing the points
    about Grids and Agents needing each other
  • Stating that the simple approach of layering is
    not enough
  • Need fine-grain intertwining for true benefits
    (stating that that is not a problem with a simple
    solution)
  • Early examples of solutions
  • Agent-based resource selection through
    re-inforcement learning
  • Reference to a paper at the same conference that
    this paper is published.
  • Second example is the use of automated
    negotiation techniques to allocated Grid
    resources
  • But the work referenced here is not about agents.
  • Not much about integration, just mention a couple
    of examples, and that this will bring challenges
    but what challenges?

16
Section 5 Robust Agile Service Oriented
Architectures
  • Draw the two we now draw the two parallel lines
    of research together to highlight their
    commonalities and complementarities.
  • Did not section 4 do that already?
  • Service as unifying concept (Autonomous services)
  • Wasnt VO the unifying concept?
  • Replicating databases and Distributed negotiation
    protocols might be used to establish the query
    throughput achievable on individual copies, such
    that community throughput is optimized.
  • Replicating databases is not so straightforward
    as the authors think. They have disregarded the
    large body of existing work on databases,
    consistency, and synchronisations.
  • The sentence on distributed negotiation protocols
    for querying databases seems unrealistic
  • Distributed planning and scheduling already
    exists in Grids, as is disregarded by the paper
  • The example of the database service is not
    convincing (blue sky thinking)

17
Section 5.2 Rich Service Models
  • Contribution of Grid is a robust lifetime and
    naming model for dynamic services
  • Service failure and scalability to benefit agent
  • But the paper mentioned before that one of the
    weakness of Grids is inability to deal with
    failure
  • Statefulness of Grids (persistency of states
    over lifetimes)
  • In contrast, agent systems address semantics but
    do not provide a consistent state model.
  • This statement should be explained, may be with
    an example and justified
  • No real solution in this section of how to
    combine the semantics of agent with state of
    services
  • Different types of semantics in agent and in Grids

18
Negotiation and Service Contracts
  • Negotiation of level of service provision
  • To establish service contracts
  • Well explained section with example of
    negotiation issues e.g. load, identity, etc.
  • What is a Byzantize failure model no reference
    here
  • Grids has a promising approach of having shared
    policy statement to represent an agreement
  • Too brief mention of strategy

19
Virtual Organisation Management
  • VOs in both Grids and Agents
  • VO creation and operation
  • Involves negotiation
  • Much work on VO in agents
  • However no reference, examples given
  • Given importance of VOs given earlier in paper,
    this section is too brief

20
Authentication, Trust and Policy
  • First paragraph from a Grid point of view,
    association of identity and community based
    authorisation
  • Second and third paragraphs Trust, negotiate
    trust and policy
  • Why is trust mentioned here?
  • Claims about expecting Grids to use agent trust
    and reputation is not justified
  • Not coherent with first paragraph
  • Trust is given too much importance with respect
    to existing work in agent community and with
    respect to the needs of the Grid
  • In fact after 3 years, scarcely any work on Grid
    and trust exists i.e. trust is not currently
    important for Grid

21
Ten Research Problems
  • Some kind of proposal for a solution towards
    Agents and Grid integration for future work
  • Remember this is a position paper not an
    implementation/theoretical paper.
  • After 3 years, which of these research problems
    are being tackled?
  • Integrated service architecture
  • Interesting, but does not seem to work after 3
    years. Few agent grid service architectures
    proposed and even less adopted
  • Trust negotiation and management
  • Again why is trust given so much importance?
  • Few Grids are concerned with trust negotiation
  • System management and troubleshooting
  • Where is the contribution of agents mentioned
    here?

22
Ten Research Problems
  • Negotiation
  • Mentioned too briefly?
  • Service composition, VO formation and management,
    System predictability
  • Seems promising but again how can agents help
    here/integrate here? No mention of agents
  • Human-computer collaboration
  • Where is the agent and where is the Grid?
  • Evaluation
  • Promising and well-explained how agent Grid can
    integrate to solve this problem
  • Semantic integration
  • Where is the Grid and agent here?

23
Critical Overview
  • First part of paper well described
  • Latter part of paper seems to be written by
    different authors and not coherent with each
    other
  • Ten research problems
  • Most written by Grid person, a couple on trust by
    agent person
  • No coherence
  • No mention of how agent Grid integration can help
    in most of these research problems
  • Conclusions is the research challenge problems

24
Questions
About PowerShow.com