SLIDE 1
Multi-Agent Systems
Jörg Denzinger
- 4. Learning in MAS
Multi-Agent Systems
Jörg Denzinger
What is learning (II)
An AI definition:
- P. Langley: Elements of Machine Learning, Morgan
4. Learning in MAS 4.1. What is learning? One definition: Bower, - - PowerPoint PPT Presentation
4. Learning in MAS 4.1. What is learning? One definition: Bower, - - PowerPoint PPT Presentation
4. Learning in MAS 4.1. What is learning? One definition: Bower, Hilgard: Theories of learning, Prentice-Hall, 1975: Learning refers to the change in a subjects behavior to a given situation brought about by his repeated experiences in
SLIDE 2
SLIDE 3
Multi-Agent Systems
Jörg Denzinger
Consequences for our agent definition
n For an agent Ag we can have fAg(s) at time t1 ≠ fAg(s) at time t2 n For a not-learning agent Ag, its learning variant AgL and a sequence of situations s1,…,si,..., we have there is an i such that Ag(sj) = AgL(sj) for j < i and Ag(si) ≠ AgL(si) n Predicting what a learning agent is doing is very difficult (see Asimovian Agents)
SLIDE 4
Multi-Agent Systems
Jörg Denzinger
A basic learning model
Langley (1996)
performance element environment knowledge base learner
SLIDE 5
Multi-Agent Systems
Jörg Denzinger
Agents and the basic learning model
n Online Learning n Offline Learning
performance element environment knowledge base learner performance element environment knowledge base learner
L L
SLIDE 6
Multi-Agent Systems
Jörg Denzinger
Learning in AI: Characterizing Learning (I)
n Learn activity and frequency: always learning vs learn phases The later allows for offline learning! n Feedback:
l Unsupervised learning (self-organized)
? ab abab ! (ab)*
SLIDE 7
Multi-Agent Systems
Jörg Denzinger
Learning in AI: Characterizing Learning (II)
n Feedback:
l Supervised learning nTeacher specifies desired system behavior nTeacher specifies quality of the performed
system behavior ababac ! ababac ababac 42
SLIDE 8
Multi-Agent Systems
Jörg Denzinger
Learning in AI: Characterizing Learning (III)
n Learn methodology
l Learning by heart l Learning by instruction and advice
acbcac acbcac, acbcac, ... ababa_ ? b
SLIDE 9
Multi-Agent Systems
Jörg Denzinger
Learning in AI: Characterizing Learning (IV)
n Learn methodology
l Learning from examples and by experiences l Learning by using analogy
Red! ! ?
SLIDE 10
Multi-Agent Systems
Jörg Denzinger
Learning in AI: Characterizing Learning (IV)
n Learn methodology
l Learning by discovery
F = G⋅ m1⋅ m2 / r2
SLIDE 11
Multi-Agent Systems
Jörg Denzinger
Learning in AI: Characterizing Learning (V)
n Most common forms of inferences Deduction: P(a); "x.(P(x) Æ Q(x)) fi Q(a) Induction: P(0); P(1)Ÿ...ŸP(n) fi "x.P(x) Abduction: P(a); "x.(Q(x) Æ P(x)) fi Q(a)
SLIDE 12
Multi-Agent Systems
Jörg Denzinger
Learning in AI: Research Areas (I)
n Neural Learning
l Simple interacting processing units l Parallel, distributed, subsymbolic
n Centralized Cognitive Learning
l Single, complex systems l Knowledge-based, symbol oriented l Especially: nRule-based Learning nInductive Learning nCase-based Reasoning
SLIDE 13
Multi-Agent Systems
Jörg Denzinger
Learning in AI: Research Areas (II)
n Distributed Cognitive Learning
l Interacting complex systems l Knowledge-based, symbol oriented l See later
n Evolutionary Learning
l Populations of (simple or complex) systems l Numerical and using random influences l Genetic Algorithms, classifier systems, etc.
SLIDE 14
Multi-Agent Systems
Jörg Denzinger
Learning in AI: Basic Problems (I)
n Similarity measures for
l Analogical reasoning l Case-based reasoning l Situations l Examples
n Generalizations for
l Decision trees l Classification tasks
SLIDE 15
Multi-Agent Systems
Jörg Denzinger