AgentOverview

• Background: discrete maths. Read chapter 4 (Mathematical Data Types) at [4]. You can skip section 4.5. Equivalently, you can read this mini-book [5], and skip sections 3.2 and 3.3.

• AgentDefinition
  • Wooldridge chap 1

• General Agent architectures
  • environments, formalization of agents and runs, state-based agents, Wooldridge chap 2 and R&N chap I.2
  • an example: RoboCup
  • FiniteStateMachines, identifying states
  • utility-based agents (Wooldridge end of chap 2), other agent architectures (R&N chap I.2)

• Problem-solving agents: R&N Chap 3 (3.1, 3.2, 3.3)
  • SearchMethods
    • reaching goals with CompleteSearch, R&N Section 3.4 (3.4.1, 3.4.2 (that's Dijkstra!), 3.4.3)
    • optimizing utility with IncompleteSearch, R&N Chap 4 (4.1, 4.1.1, 4.1.2)
  • Agents in non-deterministic environments: AND-OR trees, R&N Section 4.3
  • Adversarial search (such as in two-player games): R&N chap 5 (5.1, 5.2)

• Deductive agents
  • FormalLogic (see Genesereth ref), propositional logic: R & N Chap 7 (7.1, (7.2. is an optional but highly recommended example), 7.3, 7.4, 7.5 (only up to 7.5.1 incl.))
  • ExpertSystems and agent based on prop logic: R & N (7.5.3, 7.5.4, 7.7.1)
  • first-order logic: R & N chap 8 (especially 8.1.2, 8.2, 8.3) and chap9 (only 9.1), FormalLogic (see Genesereth ref chapters on relational logic)

• Means-end reasoning agents
  • AgentPlanning (STRIPS example link), R & N chap 10 (only 10.1, 10.2.1, 10.2.2)
  • non monotonicity, situation calculus

• Dealing with uncertainty:
  • ProbabilityBasics (+ follow link), R & N chap 13
  • BayesianNetworks (+ link), R & N 14.1 and 14.2, Decision networks (R & N 16.5), MarkovDecisionProcesses (R & N 17.1, 17.2)

• AgentAdaptability: generalities on machine learning and VersionSpaces, DecisionTrees (or R & N), ReinforcementLearning (R & N 21.1, 21.2)

• the group project!

• InterfaceAgents (Bradshaw): (no textbook) general architecture, use of k-nearest neighbor (Maes, me), grammatical inference (Schlimmer), Bayesian learning (spam-filtering), weighted majority, chronicles...

• BeliefDesireIntention: Wooldridge chap 17, ModalLogic, BdiArchitecture

• AgentCommunicationLanguage (Wooldridge chap 7)
  • AgentKnowledgeRepresentation: R & N 12.1-12.5, Wooldridge, Ontologies: description logics, SemanticWeb (Protege)
• AgentCooperation
  • Game theory (Wooldridge chap 11, R & N 17.5)
  • Mechanism design: AgentNegotiation, Auctions (Wooldridge chap 14, R & N 17.6.1)

• Multi-agent Planning (R&N 11.4)

• AgentFrameworks: Jason, JADE

• ReactiveAgents

• EthicsOfAutonomy

• AgentResearchProjects