Upload
barrie-carroll
View
216
Download
1
Tags:
Embed Size (px)
Citation preview
1 2004, G.Tecuci, Learning Agents Center
CS 7850 Fall 2004
Learning Agents Center and Computer Science Department
George Mason University
Gheorghe Tecuci [email protected]
http://lac.gmu.edu/
2 2004, G.Tecuci, Learning Agents Center
Identify a potential PhD advisor for Bob
Artificial Intelligence
Identify a potential PhD advisor for Bob in Artificial Intelligence
What is a research area of interest to Bob?
Preliminary modeling and ontology specificationPreliminary modeling and ontology specificationI need to
Therefore I need to
John Doe
Determine whether John Doe is an appropriate PhD advisor for Bob in Artificial Intelligence.
Who is a professor who is an expert in Artificial Intelligence?
Therefore I need to
continues
Dan Smith
Information security
…
…
Artificial Intelligence Information Security
PhD research area
Computer Science
John Doe Dan Smith
professor PhD advisor
PhD student
Bob
is_expert_in
student
is_interested_in
instance_of instance_of
instance_of
subconcept_of
instance_of
subconcept_of
3 2004, G.Tecuci, Learning Agents Center
Ontology specification guidelineOntology specification guideline
Concepts and instances may have properties with constant values, such as numbers (e.g. 5), strings (e.g. “yes”), or instances. The value cannot be another concept.
Whenever possible, it is recommended to use instances rather than strings, to represent knowledge in the object ontology.
Concept names are in black.Instance names are in blue.
4 2004, G.Tecuci, Learning Agents Center
Preliminary modeling and ontology specificationPreliminary modeling and ontology specification
Yes because John Doe has a tenured position
Is John Doe likely to stay on the faculty of George Mason University for the duration of Bob’s dissertation?
Yes because Jill Knox has a tenure-track position and is likely to get tenure
No because Dan Smith is likely to retire in the near future
Therefore I need to
Other possible answers
No because Amanda Rice is on a visiting position
Assess whether John Doe is an appropriate PhD Advisor for Bob in Artificial Intelligence
…
Determine whether John Doe is an appropriate PhD advisor for Bob in Artificial Intelligence
I need to
John Doe is a potential PhD Advisor for Bob in Artificial Intelligence
tenured_position
tenure_track_position
visiting_position
faculty_position
subconcept_of
long_term_position short_term_position
subconcept_of
instance_ofinstance_of
instance_of
position
retired_position
instance_of
subconcept_of
5 2004, G.Tecuci, Learning Agents Center
Ontology specificationOntology specification
PhD_student
person
faculty_member
tenured_position
tenure_track_position
visiting_position
full professor
associate_professor
assistant_professor
faculty_position
ranked_professor
studentstaff_member
subconcept_of
subconcept_of
subconcept_of
subconcept_of
subconcept_of
PhD_advisor
John Doe
instance_of
long_term_position short_term_position
has_as_position
subconcept_of
instance_of
instance_ofinstance_of
position
staff_position
subconcept_of
George Mason Universityhas_as_employer
university
University of Virginia
instance_of
employer
subconcept_of
retired_position
instance_of
Yes because John Doe has a tenured position
Is John Doe likely to stay on the faculty of George Mason University for the duration of Bob’s dissertation?
school
subconcept_of
6 2004, G.Tecuci, Learning Agents Center
Ontology specificationOntology specification
tenured_position
tenure_track_position
assistant_professor
Jill Knox
instance_ofhas_as_position
George Mason University
has_as_employeris_likely_to_obtain_position
retired_position
full_professor
Dan Smith
instance_ofhas_as_position
has_as_employer
is_likely_to_obtain_position
Yes because Jill Knox has a tenure-track position and is likely to get tenure
No because Dan Smith is likely to retire in the near future
No because Amanda Rice is on a visiting position
associate_professor
Amanda Rice
instance_ofvisiting_position
has_as_position
Is XXX likely to stay on the faculty of George Mason University for the duration of Bob’s dissertation?
7 2004, G.Tecuci, Learning Agents Center
Ontology specification: featuresOntology specification: features
is_interested_indomain
rangePhD_research_area
personis_expert_in
domain
rangePhD_research_area
person
has_as_employerdomain
rangeemployer
person
has_as_positiondomain
rangeposition
person
is_likely_to_obtain_positiondomain
rangeposition
person
8 2004, G.Tecuci, Learning Agents Center
Assess John Doe with respect to his professional reputation in Artificial Intelligence
Assess John Doe with respect to the general personality and work compatibility with Bob
Assess John Doe with respect to the expected learning experience
Assess John Doe with respect to the expected quality of student’s work
Assess John Doe with respect to the support offered to students
Assess John Doe with respect to the responsiveness to students
Assess whether John Doe is an appropriate PhD Advisor for Bob in Artificial Intelligence
Professional reputation, general personality and work compatibility, learning experience, responsiveness to the student, support, and quality of results.
Which are the main criteria to consider in assessing John Doe?
Keep the following as constants used in task names, and not as elements in the ontology:
Professional reputation, general personality and work compatibility, learning experience, responsiveness to the student, support, and quality of results.
9 2004, G.Tecuci, Learning Agents Center
Reputation among peers, expertise in Artificial Intelligence, research group status, research funding, reputation of former students, quality of PhD dissertation
Which are the main criteria to consider in judging the professional reputation of John Doe?
I need to
Assess John Doe with respect to his reputation among peers in Artificial Intelligence
Assess John Doe with respect to his expertise in Artificial Intelligence
Therefore I need to
Assess John Doe with respect to the status of his research group in Artificial Intelligence
Assess John Doe with respect to the type and level of funding of his research in Artificial Intelligence
Assess John Doe with respect to the reputation of his former students in Artificial Intelligence
Assess John Doe with respect to the expected quality of the PhD dissertation in Artificial Intelligence
Assess John Doe with respect to his professional reputation in Artificial Intelligence.
Preliminary modeling: reputationPreliminary modeling: reputation
Keep the following as constants used in task names, and not as elements in the ontology:
Reputation among peers, expertise in Artificial Intelligence, research group status, research funding, reputation of former students, quality of PhD dissertation
10 2004, G.Tecuci, Learning Agents Center
I need to
Assess John Doe with respect to his publications in Artificial Intelligence
Assess John Doe with respect to his membership in professional societies and committees related to Artificial Intelligence
Therefore I need to
Assess John Doe with respect to his research contributions in Artificial Intelligence
Research contributions, publications, and membership
Which are the main aspects that characterize reputation among peers?
Assess John Doe with respect to his reputation among peers in Artificial Intelligence
Preliminary modeling: reputation among peersPreliminary modeling: reputation among peers
Keep the following as constants used in task names, and not as elements in the ontology:
Research contributions, publications, and membership
11 2004, G.Tecuci, Learning Agents Center
I need to
Therefore I need to
Assess John Doe with respect to his research contributions in Artificial Intelligence
Doe methods for truth maintenance systems
What is a main research contribution of John Doe?
Preliminary modeling and ontology specificationPreliminary modeling and ontology specification
Assess the significance of Doe methods for truth maintenance systems
research contributionJohn Doe
instance_of
Doe methods for truth_maintenance_systems
has_as_research_contribution
contribution
subconcept_of
12 2004, G.Tecuci, Learning Agents Center
Preliminary modeling: reputation among peersPreliminary modeling: reputation among peers
Therefore I need to
citations and use in university courses
What aspects characterize the significance of a research contribution?
Assess wheather Doe methods for truth maintenance systems are cited
Assess wheather Doe methods for truth maintenance systems are used in university courses
Therefore I need to
Assess the significance of Doe methods for truth maintenance systems
Keep the following as constants used in task names, and not as elements in the ontology:
citations and use in university courses
13 2004, G.Tecuci, Learning Agents Center
I need to
Artificial Intelligence Magazine
Which is a publication that cites Doe methods for truth maintenance systems?
Preliminary modeling and ontology specificationPreliminary modeling and ontology specification
Assess wheather Doe methods for truth maintenance systems are cited
Doe methods for thruth maintenance systems are cited in Artificial Intelligence Magazine
Automated Reasoning JournalDoe methods for thruth maintenance systems are cited in Automated Reasoning Journal
Doe methods for thruth maintenance systems are cited in: Artificial Intelligence Magazine, Automated Reasoning Journal
Therefore I conclude that
Therefore I conclude that
Therefore I conclude that
research contributionJohn Doe
instance_of
Doe methods for truth_maintenance_systems
has_as_research_contribution Automated Reasoning Journal
cited_in
Artificial Intelligence Magazine
cited_in
research journal
instance_of
14 2004, G.Tecuci, Learning Agents Center
I need to
University of Illinois at Urbana Champaign
Which is a university where Doe methods for truth maintenance systems are taught?
Preliminary modeling: reputation among peersPreliminary modeling: reputation among peers
Doe methods for thruth maintenance systems are taught at University of Illinois at Urbana Champaign
University of CalgaryDoe methods for thruth maintenance systems are taught at University of Calgary
Doe methods for thruth maintenance systems are taught at: University of Illinois at Urbana Champaign, University of Calgary, University of Montreal
Assess wheather Doe methods for truth maintenance systems are used in university courses
University of Montreal Doe methods for thruth maintenance systems are taught at University of Montreal
Therefore I conclude that
Therefore I conclude that
Therefore I conclude that
Therefore I conclude that
Doe methods for truth_maintenance_systems
Univerity of Montreal
taught_at
Univerity of Illinois at Urbana Champaign
taught_at
Univerity of Calgary
taught_at
15 2004, G.Tecuci, Learning Agents Center
Ontology specificationsOntology specifications
research contributionJohn Doe
instance_of
Doe methods for truth_maintenance_systems
has_as_research_contribution
Automated Reasoning Journal
cited_in
Univerity of Montreal
taught_at
Artificial Intelligence Magazine
cited_in
Univerity of Illinois at Urbana Champaign
taught_atUniverity of Calgary
taught_at
research journal
journal
publication
instance_of
subconcept_of
subconcept_of
proceedingsbook
contribution
subconcept_of
16 2004, G.Tecuci, Learning Agents Center
Ontology specifications: featuresOntology specifications: features
has_as_research_contributiondomain
rangeresearch_contribution
person
cited_indomain
rangepublication
contribution
taught_atdomain
rangeschool
contribution
17 2004, G.Tecuci, Learning Agents Center
Preliminary modeling: reputation among peersPreliminary modeling: reputation among peers
Therefore I need to
citations and use in university courses
What aspects characterize the significance of a research contribution?
Assess wheather Doe methods for truth maintenance systems are cited
Assess wheather Doe methods for truth maintenance systems are used in university courses
Doe methods for thruth maintenance systems are cited in: Artificial Intelligence Magazine, Automated Reasoning Journal
Doe methods for thruth maintenance systems are taught at: University of Illinois at Urbana Champaign, University of Calgary, University of Montreal
Very significant
How significant are Doe methods for thruth maintenance systems?
Therefore I need to
Assess the significance of Doe methods for truth maintenance systems
Doe methods for thruth maintenance systems represent are very significant research contribution, cited in Artificial Intelligence Magazine, Automated Reasoning Journal, and taught at University of Illinois at Urbana Champaign, University of Calgary, University of Montreal.
I conclude that
Therefore I conclude that
18 2004, G.Tecuci, Learning Agents Center
I need to
Therefore I need to
Assess John Doe with respect to his research contributions in Artificial Intelligence
Doe methods for truth maintenance systems
What is a main research contribution of John Doe?
Preliminary modeling: reputation among peersPreliminary modeling: reputation among peers
Assess the significance of Doe methods for truth maintenance systems
Doe methods for thruth maintenance systems represent are very significant research contribution, cited in Artificial Intelligence Magazine, Automated Reasoning Journal, and taught at University of Illinois at Urbana Champaign, University of Calgary, University of Montreal.
John Doe is known and appreciated for his contributions in Artificial Intelligence. Doe methods for thruth maintenance systems are cited in Artificial Intelligence Magazine, Automated Reasoning Journal, and taught at University of Illinois at Urbana Champaign, University of Calgary, University of Montreal.
Yes
Is John Doe known and appreciated for his research contributions in the area of Artificial Intelligence?
Therefore I conclude that
I conclude that
19 2004, G.Tecuci, Learning Agents Center
I need to
Assess John Doe with respect to his publications in Artificial Intelligence
Assess John Doe with respect to his membership in professional societies and committees related to Artificial Intelligence
Therefore I need to
Assess John Doe with respect to his research contributions in Artificial Intelligence
Research contributions, publications, membership
Which are the main aspects that characterize reputation among peers?
Assess John Doe with respect to his reputation among peers in Artificial Intelligence
Preliminary modeling: reputation among peersPreliminary modeling: reputation among peers
John Doe …
John Doe …
… …
John Doe is known and appreciated for his contributions in Artificial Intelligence. Doe methods for thruth maintenance systems are cited in Artificial Intelligence Magazine, Automated Reasoning Journal, and taught at University of Illinois at Urbana Champaign, University of Calgary, University of Montreal.
20 2004, G.Tecuci, Learning Agents Center
I need to
Assess John Doe with respect to his publications in Artificial Intelligence
Assess John Doe with respect to his membership in professional societies and committees related to Artificial Intelligence
Therefore I need to
Assess John Doe with respect to his research contributions in Artificial Intelligence
Research contributions, publications, membership
Which are the main aspects that characterize reputation among peers?
Assess John Doe with respect to his reputation among peers in Artificial Intelligence
Preliminary modeling: reputation among peersPreliminary modeling: reputation among peers
John Doe …
John Doe …
… …
John Doe is known and appreciated for his contributions in Artificial Intelligence. Doe methods for thruth maintenance systems are cited in Artificial Intelligence Magazine, Automated Reasoning Journal, and taught at University of Illinois at Urbana Champaign, University of Calgary, University of Montreal.
John Doe has an excellent reputation
How good is the reputation of John Doe among his peers?
I determined that
John Doe has an excellent reputation among peers in Artificial Intelligence.
I conclude that
21 2004, G.Tecuci, Learning Agents Center
Reputation among peers, expertise in Artificial Intelligence, research group status, research funding, reputation of former students, quality of PhD dissertation
Which are the main criteria to consider in judging the professional reputation of John Doe?
I need to
Assess John Doe with respect to his reputation among peers in Artificial Intelligence
Assess John Doe with respect to his expertise in Artificial Intelligence
Therefore I need to
Assess John Doe with respect to the status of his research group in Artificial Intelligence
Assess John Doe with respect to the type and level of funding of his research in Artificial Intelligence
Assess John Doe with respect to the reputation of his former students in Artificial Intelligence
Assess John Doe with respect to the expected quality of the PhD dissertation in Artificial Intelligence
Assess John Doe with respect to his professional reputation in Artificial Intelligence.
Preliminary modeling: reputationPreliminary modeling: reputation
John Doe has an excellent reputation
How good is the reputation of John Doe in Artificial Intelligence?
John Doe …
I determined that
John Doe …
John Doe …
John Doe …
John Doe …
John Doe has an excellent reputation in Artificial Intelligence.
I conclude that
John Doe has an excellent reputation among peers ...
22 2004, G.Tecuci, Learning Agents Center
John Doe would be a very good PhD Advisor for Bob in Artificial Intelligence.
Assess John Doe with respect to his professional reputation in Artificial Intelligence
Assess John Doe with respect to the general personality and work compatibility with Bob
Assess John Doe with respect to the expected learning experience
John Doe …
Assess John Doe with respect to the expected quality of student’s work
Assess John Doe with respect to the support offered to students
Assess John Doe with respect to the responsiveness to students
Assess whether John Doe is an appropriate PhD Advisor for Bob in Artificial Intelligence
How good a PhD Advisor would John Doe be for Bob in Artificial Intelligence?
Very good
Professional reputation, general personality and work compatibility, learning experience, responsiveness to the student, support, and quality of results.
Which are the main criteria to consider in assessing John Doe?
John Doe has an excellent reputation in Artificial Intelligence.
John Doe …
John Doe …
John Doe …
John Doe …
23 2004, G.Tecuci, Learning Agents Center
Preliminary modeling: necessary conditionsPreliminary modeling: necessary conditions
Yes because John Doe is a tenured faculty
Is John Doe likely to stay on the faculty for the duration of Bob’s dissertation?
Yes because X has a very good record and is very likely to get tenure
No because X is likely to retire in the near future
Therefore I need to
Other possible answers
No because X is a visiting faculty that will leave next year
…
Determine whether John Doe is an appropriate PhD advisor for Bob in Artificial Intelligence
I need to
John Doe would be a very good PhD Advisor for Bob in Artificial Intelligence.
John Doe is a potential PhD Advisor for Bob in Artificial Intelligence.
John Doe would be a very good PhD Advisor for Bob in Artificial Intelligence.
Assess whether John Doe is an appropriate PhD Advisor for Bob in Artificial Intelligence
John Doe is a potential PhD Advisor for Bob in Artificial Intelligence
24 2004, G.Tecuci, Learning Agents Center
Identify a potential PhD advisor for Bob
Artificial Intelligence
Identify a potential PhD advisor for Bob in Artificial Intelligence
What is a research area of interest to Bob?
Preliminary modelingPreliminary modelingI need to
Therefore I need to
John Doe
Determine whether John Doe is an appropriate PhD advisor for Bob in Artificial Intelligence.
Who is a professor who is an expert in Artificial Intelligence?
Therefore I need to
continues
Dan Smith
Information security
…
…
John Doe would be a very good PhD Advisor for Bob in Artificial Intelligence.
John Doe would be a very good PhD Advisor for Bob.
John Doe would be a very good PhD Advisor for Bob in Artificial Intelligence.
25 2004, G.Tecuci, Learning Agents Center
Individual assignmentIndividual assignment
• Install Disciple-RKF on your computer.
• Introduce the folder “PhD-Adv” (received by email from the instructor) in the folder “\Disciple-RKF\Repository”
• Define the ontology specified in these notes by using the ontology development tools of Disciple.
• Introduce the modeling specified in these notes (the viewgraphs 2 to12 only) by using the modeling assistant of Disciple.
Due date: October 28th, 2004