CogSci 207 Midterm Review Fall 2004 Praveen Paritosh 10/18/04
Where are we? Minsky/Turing/Miller –Computation and Cognition Markman/Lenat/Cyc –Representation Forbus/Bredeweg/Vm odel –Qualitative Reasoning Riesbeck/Lee/Batali –Natural Language Emotions Analogy and Similarity Learning/Education Consciousness Now Coming
Cognitive modeling Psychology – mind Artificial Intelligence – computer Neuroscience – brain Philosophy – everything Linguistics – language Cognitive Science – builds up on all of the above. Goal: to understand cognition by building computational models that operate on representations.
CogSci 207: A slice of Cognitive Modeling What are computational models and what are they capable of? –Dispel some myths: Turing/Minsky –Show some examples of models of cognitive phenomena: Qualitative Reasoning/NLP –More to come: Analogy/Emotions/Learning What are representations and what can we do with them? –Foundations: Markman/Lenat/Minsky –Some examples: Cyc/QP/Episodic –More to come: Analogy/Emotions/Education
Computational Models Computers can do more than just what we tell them. Programs can encode library of methods/ strategies/ plans/ heuristics* and the composition of these can be completely novel and unanticipated. Programs can reflect about their own progress and decide on courses of action. * The Art Of War/ The Prince/ Andy Gordon’s work on strategies, for example.
Representations Cyc –Predicate calculus based –Organizing principles key Mereology Ontology Domain specific knowledge Microtheories QP –Strong organizing principles –Built upon structure of everyday physical world
Questions?