Métodos de Computação Inteligentes 2005.1
General information:
§
Lecturer:
Jacques Robin
§ Teaching assistants: Fábio Moura, Jairson Vitorino & Marco Aurélio da Silva
§
When: 3a
§ Where: M1
§ This page: www.cin.ufpe.br/~in1006/2005/
§
Newsgroup: depto.courses.posgrad.in1006
Bibliography:
§ Artificial Intelligence a Modern Approach (2nd Ed.), S. Russell & P. Norvig, 2002, Prentice-Hall. Site: aima.cs.berkeley.edu
§ Introduction to Multi-Agent Systems, M. Wooldridge, 2002, Wiley. Site: www.csc.liv.ac.uk/~mjw/pubs/imas/
§
UML 2 Toolkit. Eriksson, H.E., Penker, M.,
§ The Object Constraint Language: Getting Your Models Ready for MDA (2nd Ed.) Warmer, J. & Kleppe, A. 2003. Addison-Wesley.
§ Logic, Programming and Prolog (2nd Ed). Nilsson, U. & Maluszynski. 2000. Site: http://www.ida.liu.se/~ulfni/lpp/
§ Constraint Programming: an Introduction. Marriott, K. & Stuckey, P. 1998. MIT Press.
§ Essentials of Constraint Programming. Frühwirth, T. & Abdennhader, S. 2003. Springer.
§ Constraint Processing. Dechter, R. Morgan-Kaufmann. 2003.
§ Ontologies: a Silver Bullet for Knowledge Management and Electronic Commerce. Fensel, D. 2003. Springer.
Evaluation:
§
Individual
multiple-answer final exam: 30% of the grade
§
Oral seminar: 20% of
the grade
§
Group project: 50%
of the grade
§
10% for each stage
deliverable
§
10% for final
revised and assembled deliverables
§
Each deliverable:
§
6% for artifact
(model or code) quality
§ 4% for report quality
Grades breakdown:
§
Oral seminar:
§
Meeting first
advising deadline: 1 point
§
Conciseness: 1 point
§
Breadth
completeness: 1 point
§
Depth completeness:
1 point
§
Correctness: 1 point
§
Command/understanding
of material: 1 point
§
Clarity: 1 point
§
Examples: 1 point
§
Oral skills: 1 point
§
Questions and
interactions with students: 1 point
§
Project deliverable
presentation:
§
Presentation itself:
§
Command/understanding
of material: 1 point
§
Design choice
motivation: 1 point
§
Clarity: 1 point
§
Oral skills: 1 point
§
For Models:
§
Breath completeness:
1 point
§
Depth completeness:
1 point
§
Correctness: 1 point
§
Design quality: 1
point
§
Conciseness: 1 point
§
Explanations in written
report: 1 point
§
For Code:
§
Breadth of
functionalities: 1 point
§
Functional tests and
robustness: 1 point
§
Performance tests
and efficiency: 1 point
§
Interface and
user-friendliness: 1 point
§
Modularity and
reusability: 1 point
§
Documentation and
comments: 1 point
Seminar advising:
§
Two compulsory
meeting with adviser: the first two weeks
before seminar’s date, the second one week
before.
§
For the first
meeting, the student have ready a detailed outline of the seminar that
indicates the title and planned content of each slide
§
For the second
meeting, the student must have the presentation ready to rehearse it with the
adviser
§
Failure to meet the
first deadline will result in loosing one points from the seminar’s grade
§ Failure to meet the second deadline will result in the cancellation of the seminar, with the adviser presenting the lecture and the student getting the grade zero for the seminar
Roster:
|
Aluno |
Pesquisa |
Conhecimento Prévio |
||||||||
|
Login |
Curso |
Área |
Orientador |
Tema |
Java |
UML |
OCL |
Lógica |
Disciplinas de IA já cursadas
|
|
Alexandre Maciel |
amam |
M |
Comp. Int. |
Edson Carvalho |
|
+ |
+ |
- |
+ |
IA |
|
Alexsandro Farias |
ajmf |
M |
Comp. Int. |
Jacques Robin |
|
+ |
+ |
- |
+ |
IA |
|
Daniela Cargnin |
dc2 |
M |
Comp. Int. |
Germano Vasconcelos |
|
- |
+ |
- |
+ |
IA |
|
Eduardo Dominoni |
ecgd |
M |
Agentes Autônomos |
Patrícia Tedesco |
|
+ |
+ |
- |
+ |
Sistemas
Inteligentes, Agentes Autônomos, |
|
Gláucya Boechat |
gcb |
M |
Comp. Int. |
Edson Carvalho |
|
+ |
+ |
- |
+ |
IA |
|
Humberto Brandão |
hcbo |
M |
|
Germano Vasconcelos |
|
+ |
+ |
- |
+ |
IA, Otimização
Comb., Recup. Informação |
|
Jeneffer Ferreira |
jcf |
M |
Comp. Int. |
Edson Carvalho |
|
+ |
+ |
- |
+ |
IA |
|
João Farias |
jpf2 |
Isolada |
|
|
|
+ |
+ |
- |
+ |
IA |
|
João Paulo Rolim |
jpcr |
M |
Arquitetura de Comp. |
Sérgio Calvacante |
|
+ |
+ |
- |
+ |
Fundamento de |
|
José Alexandrino |
jla |
M |
Comp. Int. |
Edson Carvalho |
|
+ |
+ |
- |
+ |
IA |
|
Leandro Almeida |
lma3 |
M |
Comp. Int. |
Teresa Ludermir |
|
+ |
+ |
- |
+ |
IA |
|
Luiz Francisco Lacerda |
lfblj |
M |
Comp. Int. |
Jacques Robin |
Refactoring do Replay – MaracatuRFC |
+ |
+ |
- |
+/- |
Fundamentos de |
|
Luiz Gustavo Carvalho |
lgcc |
M |
Comp. Int. |
Edson Carvalho |
|
+ |
+ |
- |
+ |
IA |
|
Marcio Carvalho |
mrl2 |
M |
Comp. Int. |
Teresa Ludermir |
Otimização de Redes Neurais |
+ |
+ |
- |
+ |
IA |
|
Tarcisio Gurgel |
tbg |
M |
Mineração de Dados |
Paulo Adeodato |
Mineração de Dados Aplicada
à Epidemiologia |
+ |
+ |
- |
+ |
Fundamento de IA |
Unit 1: Introduction to Intelligent Agents
Lecture 0: Course Overview 19/04
§ Jairson
§ Outline: this page
Lecture 1: Artificial Intelligence and Intelligent Agents 26/04
§ Jairson
§
§ Sections 1, 2.1, 2.2, 7.2 of Russell & Norvig
§ Chapters 1, 2 of Wooldridge
§ Outline:
§ What is artificial intelligence?
§ What is an agent?
§ What is an intelligent agent?
§ Applications of intelligent agents
§
Slides: IntelligentAgents.ppt
Lecture 2: Agent Environments and Architectures 28/04
§ Fábio
§
§ Sections 2.3-2.4 of Russell & Norvig
§ Chapter 5 of Wooldridge
§ Outline:
§ Classifying dimensions of agent environments
§ Internal architectures of agents
§ Slides: AgentEnvironmentsArchitectures.ppt
Lecture 3: Knowledge-Based Agents 03/05
§ Jairson
§
§ Sections 7.1, 8.4 & 10.1 of Russell & Norvig
§ Learning, Bayesian Probability, Graphical Models and Abduction: http://www.cs.ubc.ca/spider/poole/papers/indab.pdf
§ Outline:
§ Inference engines and declarative knowledge bases
§ Classifying dimensions of knowledge base elements
§ Commitments of knowledge representation languages
§ Automated reasoning tasks
§ Monotonic deduction
§ Belief revision
§ Constraint solving
§ Optimization
§ Abduction
§ Inheritance
§ Induction
§ Analogy
§ Internal architecture of knowledge-based agents
§ Knowledge acquisition
§ Slides: KnowledgeBasedAgent.ppt
Unit 2: Search
Lecture 4: Problem Solving through
Search 05/05
§
João Fárias advised by Jairson
§
§
Outline:
§
Agent reasoning as
navigating a space of possibilities
§
Exhaustive search
strategies
§
Avoiding repeated
states
§
Searching with
partial information
§ Slides: AIasSearch.ppt
Lecture 5: Heuristic Search Algorithms 10/05
§
Humberto César Brandão
advised by Jairson
§
§
Outline:
§
Heuristic global
search
§
Domain-dependent
heuristic function design
§
Local search
§
Online search
Slides: HeuristicSearch.ppt
Lecture 6: Constraint Satisfaction Search 12/05
§ Leandro Maciel advised by Jairson
§
§ Section 5 of Russell & Norvig
§ Sections 6.1, 6.2 and 6.4 of Dechter
§ Outline:
§ Constraint Satisfaction Problems
§ Exhaustive Global CSP search
§ Domain-independent heuristics for global CSP search
§ Local CSP search
Slides: CSP.ppt
Unit 3: Knowledge Representation
Lecture 7: Object-Oriented Knowledge Representation 17/05
§ Fábio
§
§ Sections 10.2, 10.6 of Russell & Norvig
§ Sections 1, 2, 4 and pp. 145-172 of Eriksson & al.
§ Outline:
§ Overview of object-oriented knowledge representation
§ Semantic networks
§ Frames
§ Overview of UML
§ UML class diagrams
§ UML object diagrams
§ UML activity diagrams
§
Slides: OOKRUML.ppt
Lecture 8: Ontologies and Logical Constraints on Object-Oriented Knowledge 19/05
§ Fábio
§
§ Section 10.1 of Russell & Norvig
§ Section 2 of Fensel
§ Sections 1.3-1.5, 3.1-3.3, 3.6, 6.1-6.5, 6.7, 7-10 of Warmer & Kleppe
§ Outline:
§ What is an ontology?
§ Minimal definition
§ Purposes and origins of ontologies
§ Elements and diversity of ontologies
§ Overview of OCL
§ OCL to adorn UML class diagrams
§ OCL expressions
§ OCL basic types
§ OCL enumerations and collections
§ OCL to adorn UML activity diagrams
§
Slides: OntologiesOCL.ppt
Lecture 9: UML and OCL Knowledge
Representation Tools 24/05
§
Fábio
§
Outline:
§
Class diagrams with
Rational Rose
§
Activity diagrams
with Rational Rose
§
OCL constraints with
Poseidon
§
Slides: UMLOCLTools.ppt
Lecture 10: Project Topics 31/05
§
Jairson
§
Outline:
§
Divide students in two
halves: one for the programming project and one for the modeling project
§
Programming project:
one large team
§
Modeling project:
two teams of equal size
§ Modeling teams project: UML and OCL ontology of selected topics presented in lectures and in the reading material
§ Programming team project: Java implementation of a selected topics presented in lectures and in the reading material (fewer topics than for the modeling teams project)
§
Both projects:
§ Divided in four stages
§ Partial deliverable at the end of each stage
§ Final overall result correcting each deliverable and assembling them together
§
Slides: ProjectTopics.ppt
Unit 4: Logical Knowledge Formalization and Automated Reasoning
Lecture 11: Rule-Based Constraint Programming 02/06
§ Jairson
§
§ Chapter 7 of Frühwirth & Abdennadher
§ Chapters 2 of Marriott & Stuckey
§ Chapter 5 of Rule-Based Constraint Programming: Theory and Practice: http://www.cs.guc.edu.eg/faculty/sabdennadher/Publikationen/habil-schriftNew.ps.gz
§ Outline:
§ Limitations of CSP
§ Simplification, Propagation, Optimization and Implication
§ Overview of Constraint Handling Rules (CHR)
§ CHR syntax
§ CHR declarative logical semantics
§ CHR operational semantics
§ CHRD: extension with disjunctive bodies
§ Advantages and limitation of CHRD as a knowledge representation language
§
Slides: CHR.ppt
Lecture 12: CHREK: a Java Rule-Based Constraint
Programming Platform 07/06
§
Jairson & Marco Aurélio
§
§
Slides: CHREK.ppt
Lecture 13: Revision 14/06
§ Jacques
Lecture 14: Theorem Proving 23/06, 30/06
§ Jacques
§
§ Sections 7.2-7.7, 8, 9.1-9.2, 9.5 of Russell & Norvig
§ Outline:
§ Propositional logic syntax
§ Propositional logic semantics
§ Theorem proving using propositional logic
§ Advantages and limitations of propositional logic as a knowledge representation language
§ First-order logic syntax
§ First-order logic semantics
§ Unification
§ Theorem proving using full first-order logic
§ Advantages and limitations of full first-order logic as a knowledge representation language
§ Comparison with CHRD
§
Slides: TheoremProving.ppt
Lecture 15: Tabled Monotonic Logic Programming 30/06, 05/07
§ Jeneffer Cristine Ferreira & Glaucya Carreiro Boechat advised by Jacques
§
§ Sections 9.3-9.4 of Russell & Norvig
§ Sections 2 and 3 (skip the proofs) of Nilsson & Maluszynski
§ Section 3.1 of Tabled Evaluation with Delaying for General Logic Programs:
§ Sections 2.1, 2.2, 3.3, 4.1-4.2, 4.4 and 5 of HiLog: a Foundation for High-Order Logic Programming:
§ Outline:
§ The metaphors of logic programming
§ Definite logic programs (DLP) syntax
§ DLP declarative semantics:
§ Closed-World Hypothesis
§
Intentional DLP declarative semantics:
§ Extensional DLP declarative semantics: least Herbrand model
§ DLP operational semantics:
§ SLD resolution for DLP (goal-driven backward chaining)
§ Immediate consequence operator fixed point (data driven forward chaining)
§ SLG resolution for DLP (both goal and data driven backward chaining)
§ Advantages and limitation of DLP as a knowledge representation language
§ HiLog
§ Comparison with CHRD
§ Comparison with full first-order logic
§ Slides: TMLP.ppt
Lecture 16: Reasoning about Actions and Change 05/07
§ Marcio Ribeiro de Carvalho advised by Jacques
§
§ Sections 10.3, 10.4, 10.7 and 10.8 of Russell & Norvig
§
Sections 2, 7 of Transaction
Logic Programming, A Logic of Procedural and Declarative Knowledge: http://citeseer.ist.psu.edu/10586.html
§ The Event Calculus Explained: http://casbah.ee.ic.ac.uk/~mpsha/ECExplained.pdf
§ Outline:
§ The frame, qualification and ramification problem
§ Situation calculus
§ Event calculus
§ Transaction logic
§ Belief revision and truth-maintenance systems
§ Comparison
§
Slides: ActionsChange.ppt
Lecture 17: Constraint Logic Programming 07/07, 14/07
§ José Lima Alexandrino advised by Jacques
§
§ p. 294 of Russel & Norvig
§ Chapters 4, 7 of Marriott & Stuckey
§ Outline:
§ Limitations of CSP
§ Limitations of Prolog
§ Goal and constraint evaluation in CLP
§ CLP as a reasoning service family
§
Slides: CLP.ppt
Deadline for
First Project Deliverable:
§ Modeling Teams: UML/OCL
Ontology of CSP Problems and Algorithms
§ Programming Teams: Java Implementation of Conflict-Directed Backjumping Search for Finite Domain Constraint Satisfaction
Lecture 18: Presentation, Feedback and Discussion 1st
Project Deliverable 26/07
§
Jacques
§
Orientação: seminário de Daniela e Tarcisio
Lecture 19: Abduction and Negation as Failure 28/07
§ Jacques
§
§ Sections 10.7 of Russell & Norvig
§ Section 4 (skip the proofs) of Nilsson & Maluszynski
§ Sections 1-4,8 of The Role of Abduction in Logic Programming: http://citeseer.ist.psu.edu/kakas98role.html
§ Outline:
§ Limitations of monotonic reasoning with incomplete knowledge
§ Negation as failure in Prolog: negative hypothetical reasoning with knowledge gap
§
§ SLDNF
§ Well-founded models
§ Answer set programming
§ Default logic
§ Abductive frameworks: positive hypothetical reasoning with knowledge gap
§ Applications of abduction
§ Abduction and default logic
§ Abduction and negation as failure
§ Abduction, belief revision and truth-maintenance
§
Slides: AbductionNAF.ppt
Lecture 20: Object-Oriented Rule-Based Programming 28/07
§ Luiz Francisco Lacerda advised by Jacques
§
§ Sections 1-4 of Logical Foundations of Object-Oriented and Frame-Based Languages: ftp://ftp.cs.sunysb.edu/pub/TechReports/kifer/flogic.pdf
§ Chapter 2 of A Model Theory for Non-Monotonic Multiple Value and Code Inheritance in Object-Oriented Knowledge Base: http://www.cse.buffalo.edu/faculty/gzyang/papers/yangPhDdissertation.pdf
§ Sections 6, 7 of Flora-2: User’s Manual: http://flora.sourceforge.net/docs/floraManual.pdf
§ ILOG JRules Whitepaper: http://www.ilog.com/products/rules/whitepapers/index.cfm
§ Outline:
§ Limitations of Object-Oriented Programming for AI
§ Limitations of Logic, Constraint and Rule-Based Programming for AI
§ Embedding rules into objects
§ Embedding objects into rules
§ JRules: object-oriented production system
§ Frame Logic: object-oriented logic programming
§ Structural and behavioral inheritance
§ Value and code inheritance
§ Monotonic and non-monotonic inheritance
§ Interaction between inheritance and deduction
§
Slides: OOLP.ppt
§
Orientação: seminário
de Eduardo e projetos
Lecture 21: XSB and Flora: a Versatile Logic Programming
Platform 30/07
§
Marco Aurélio & Jacques
§
§
The XSB Manual
§
The Flora-2 Manual
§
Outline:
§ Starting XSB
§ Loading and Compiling an XSB program
§ Compiling an XSB program
§ Submitting XSB queries
§ Tracing an XSB query
§ Debugging an XSB programs
§ Pitfalls in XSB programming
§ Starting Flora
§ Loading and Compiling a Flora program
§ Submitting Flora queries
§ Tracing a Flora query
§ Debugging a monotonic Flora program
§ Debugging a Flora program with updates
§ Interaction between updates and tabling
§ Pitfalls of Flora programming
§
Slides: XSBFlora.ppt
Lecture 22: Description Logics and the Semantic Web 02/08
§
Daniela Cargnin advised by Jacques
§
§
pp. 353-354 do Russell & Norvig
§
An
introduction to description logics: http://www.inf.unibz.it/~franconi/dl/course/dlhb/dlhb-01.pdf
§
pp.
142-148, 154-158, 172-183 of Description
logics: comparison with other formalisms
§
Chapters 1-4 of A
semantic web primer: http://bbs.sjtu.edu.cn/file/SemanticWeb/1096520972228120.pdf
§
Outline:
§
Description logics
representation languages
§
Reasoning services
of inference engines for description logics
§
Comparison between
description logics and transaction frame logic
§
The semantic web
vision
§
XML
§
RDF
§
RDFS
§
OWL
§
Inference engines
for OWL
§
Semantic web
ontology editors
§
Slides: DescritionLogicsSemanticWeb.ppt
§
Orientação: seminário de Alexandre Maciel e de projetos
Unit 5: Probabilistic and Decision-Theoretic Reasoning
Lecture 23: Propositional Bayesian Reasoning 02/08
§
Tarcisio Gurgel advised by Jacques
§
§
Chapter 13 and Sections 14.1-14.5 of Russell
& Norvig
§
Outline:
§
Reasoning with uncertain knowlegde
§
Probability theory
§
Inference with full joint probability
distribution
§
Inference with Bayes’
rule
§
Bayesian networks
§
Exact inference in Bayesian networks of
discrete variables
§
Handling continuous variable in Bayesian
network inference
§
Approximate inference in Bayesian networks
§ Slides: BayesianReasoning.ppt
§
Deadline for Second Project Deliverable:
§ Modeling Teams:
UML/OCL Ontology of Monotonic Logic-Based Automated Reasoning
Programming Teams: Java/CHREK Implementation of CLP(FD)
§
Orientação: seminário de João Paulo e de Eduardo
Lecture 24: Presentation, Feedback and Discussion of 2nd
Project Deliverable 04/08
§
Jacques and Jairson
§
Orientação: de projeto
§
Lecture 25: Propositional Decision-Theoretic Reasoning 04/08
§
Eduardo Dominoni advised by Jacques
§
§
Chapters 16, 17 of Russell & Norvig
§
Outline:
§
From pursing a single goal to trade-off and
weight multiple goals
§
Utility theory and one-shot decision problems
§
Decision networks
§
The utility of information
§
Sequential decision problems
§
Policy iteration
§
Partially observable markov
decision processes
§
Decision theoretic reasoning
§
Slides: DecisionTheoreticReasoning.ppt
Lecture 26: First-Order Bayesian Reasoning 09/08
§
Alexandre Maciel advised by Jacques
§
§
Section 14.6 of Russell & Norvig
§
Bayesian Logic
Programs: ftp://ftp.informatik.uni-freiburg.de/documents/reports/report151/report00151.ps.gz
§
CLP(BN),
Constraint Logic Programming for Probabilistic Knowledge: http://www.cos.ufrj.br/~vitor/Yap/clpbn/
§
Probabilistic
space partitioning in constraint logic programming http://www-users.cs.york.ac.uk/~nicos/pbs/Asian04.ps.gz
§
Outline:
§ Limitations of Bayesian Networks
§ Extending Bayesian Networks with database relations
§ Extending Logic Programs with Conditional Probability Tables
§ Extending Logic Programs with Probabilistic Constraints
§ Slides: FirstOrderProbabilisticReasoning.ppt
§
Deadline for Third Project Deliverable:
§ Modeling Teams: UML/OCL
Ontology of Non-Monotonic Logic-Based Automated Reasoning
§ Programming Teams:
Java/CHREK Implementation of Object-Oriented CLP(FD)
§
Orientação: seminário de João Paulo e de projeto
Unit 6: Conclusion
Lecture 28: Artificial Intelligence Paradigms 09/08
§ João Paulo Rolim advised by Jacques
§
§ Sections 3.1, 5.1, 13.1, 14.1, 18.1-2 of Russell & Norvig
§ Sub-section 4.3 about genetic algorithms of Russell & Norvig
§ Sub-section 14.7 about fuzzy logic of Russell & Norvig
§ pp. 736-739, 744-748 of Sub-section 20.5 of Russell & Norvig
§ Section 9 of Wooldridge
§ Outline:
§ AI as search (navigation metaphor)
§ AI as symbolic processing (logic metaphor)
§ AI as numerical processing
§ Probabilistic processing
§ Fuzzy processing (linguistic metaphor)
§ AI as interaction (sociology and economics metaphor)
§ AI as network activation (neurology metaphor)
§ AI as evolution (genetic metaphor)
§ Hybrid paradigms
§
Slides: AIParadigms.ppt
Lecture 29: Final Exam 11/08
§
Jacques
Lecture 30: 18/08
§
Deadline for Fourth Project Deliverable:
§
Modeling Teams: UML/OCL Ontology of Bayesian and
Decision-Theoretic Automated Reasoning
§
Programming Teams: Java/CHREK or Flora/CHR/XSB
Implementation of Object-Oriented Bayesian Networks
Lecture 31: Presentation, Feedback and Discussion of 4th
Project Deliverable 23/08
Jairson and Fábio