Course Instructor
Pat Langley (langley@asu.edu)
Computer Science and Enginnering / Psychology
Arizona State University
Tempe, AZ 85287
Office: BYENG 446
Telephone: 480-965-8850
Teaching Assistant
Glen Hunt (glen.hunt@asu.edu)
PhD Student, Computer Science and Enginnering
Arizona State University
Tempe, AZ 85287
Office: BYENG 443 AB
Course Summary
Artificial intelligence (AI) attempts to answer one of the basic questions of science: What is the nature of the mind? It approaches this challenge from a computational perspective, examining the mental structures or representations that support intelligent behavior and the processes that operate over them to produce it.
Historically, studies of human cognition have served as the inspiration for artificial intelligence in two ways. First, they suggests distinct capacities that an intelligent system should possess, such as the ability to draw inferences, execute complex procedures, solve novel problems, and use language, as well as pose challenge problems related to these capacities. Second, AI has often incorporated insights about human cognition into the representations and processes used to construct intelligent agents.
Unfortunately, much of the recent research in artificial intelligence ignores cognitive science. The field has also become fragmented, devoting considerable attention to the components of intelligence but far less on how they fit together. In this course, we will review work that goes against these trends by making contact with results on human cognition and that AI's original dream of building complete intelligent agents.
The notion of physical symbol systems will figure prominently in the course, along with list structures, pattern matching, and rule-based systems. These ideas go back to the shared origins of AI and cognitive science in the 1950s, but they are still central to constructing intelligent agents and to modeling high-level cognition in humans. We will also draw frequently on a more recent idea — cognitive architectures — that offers a powerful approach to building intelligent systems.
Coursework will include reading and discussion of papers, combined with programming exercises in high-level languages that embody assumptions about the nature of intelligence. Participants will acquire expertise in designing, implementing, and running intelligent systems, along with understanding how the underlying technology draws on our knowledge of human and machine cognition.
Course Logistics
- Time: Mondays and Wednesdays, 5:00PM — 6:15PM
- Location: BYAC 270
- Office Hours: BYENG 446, Mondays 6:15PM — 7:15PM or by appointment
Prerequisites
Course prerequisites include a basic familiarity with concepts about data structures and processes from computer science. Students should have some programming experience, but exercises will use high-level languages for symbolic processing that should be easier to learn than traditional programming formalisms. However, participants should be able to think computationally in terms of symbolic or logical knowledge structures and the mechanisms that operate on them. Most important, they should be interested in the processes that underlie intelligence in humans and machines.
Assignments and Grading
Most meetings will combine a brief lecture about key ideas on some aspects of cognitive systems with group discussion of papers from the literature on those topics. To ensure that graduate participants have digested the material, they should bring one question that arose from their reading of each assigned paper, which we will use to direct discussion.
Participants will also complete a number of exercises that involve building cognitive systems of increasing complexity, which they should turn in at the beginning of the meeting for which they are due. Each student will also have a total of four free late days to use as he sees fit on the assignments. Once these late days are exhausted, we will penalize any exercises turned after its due date at the rate of 20 percent per calendar day late (or fraction thereof).
Undergraduate participants' grades will be based on correctness of their answers to exercises (50 percent), on the course project (25 percent), and on the course final (25 percent).
Graduate participants' grades will be based on correctness of their answers to exercises (40 percent), on the course project (20 percent), and on the course final (20 percent), along with contributions to class discussion (10 percent) and the quality of a written research proposal (10 percent).
Course Readings
The recommended textbooks for this course appear below, but these are mainly to provide background material. The regular reading assignments will come from the broader literature in artificial intelligence and cognitive science. We will provide URLs for papers on the course Web site.
- Nilsson, N. J. (1998). Artificial intelligence: A new synthesis. San Francisco: Morgan Kaufmann.
- Anderson, J. R. (2009). Cognitive psychology and its implications. New York: Worth Publishing.