Professor Matt Jones (mcj@colorado.edu)
Department of Psychology
Muenzinger E229
Phone TBA
Office Hours:  W 13:30-15:30 or by appointment

Professor Michael Mozer (mozer@colorado.edu)
Department of Computer Science
Engineering Center Office Tower 7-41
(303) 492-4103
Office Hours:  Tu 12:30-13:30, W 13:00-14:00, or by appointment

Cognitive modeling involves the design of computer simulation and mathematical models of human cognition and perception.  The goals of cognitive modeling include:

• understanding mechanisms of information processing in the human brain
• interpreting behavioral, neuropsychological, and neuroscientific data
• suggesting techniques for remediation of cognitive deficits due to brain injury and developmental disorders
• suggesting techniques for facilitating learning in normal cognition
• constructing computer architectures to mimic human-like intelligence.

The range of  modeling tools in cogntiive science are vast, and include production systems (sequential rule fiiring), neural networks, Bayesian probabilistic models, and pure mathematical theories.  All of these tools share the following virtues:

• Models force you to be explicit about your hypotheses and assumptions.
• Models provide a framework for integrating knowledge from various fields.
• Models allow you to bserve complex interactions among hypotheses.
• Models provide the ultimate in controlled experimentation.
• Models lead to empirical predictions.
• Models provide the sort of mechanistic framework that will ultimately be required in a theory of cortical computation.

Models can be built at many levels of the reductionist hierarchy.  Single cell models characterize the details of neural function:  ion flow, membrane depolarization, neurotransmitter release, action potentials, neuromodulatory interactions.  Network models focus on neurophysiology and neuroanatomy of cortical regions, cell firing patterns, inhibitory interactions, and neural mechanisms of learning.. Functional models characterize the operation and interaction of components of the cognitive architecture and emphasize the transformation of representations.  Finally, models at the computational level focus on the input-output behavior of the system and provide a mathematical characterization of cognition and learning.  In this seminar, we'll emphasize the functional and computational level models.  Randy O'Reilly and Yuko Munakata in psychology teach an outstanding course that focuses on the single-cell and network levels.

We will read state-of-the-art research in the field of cognitive modeling, critique the work, and discuss its contributions to the field. Students will have the opportunity to develop their own models as well.  The course participants are likely to be a diverse group of students and faculty, some with primarily an engineering/CS focus and others primarily interested in cognitive science and cognitive neuroscience.

In 2008, we plan to focus on sequential dependencies in human cognition, i.e., how one experience influences subsequent perceptions, decisions, and judgements.  As a trivial example, if I ask the following question: "On a 1 to 10 scale, how bad is it to steal from a homeless person?", your response will depend on the preceding question I've asked.  Stealing will be given a lower rating if the previous question is, "How bad is it to shoot at someone who annoys you?" than if the previous question is, "How bad is it to not leave a 10% tip?

The instructors believe that sequential dependencies offer deep insight into mechanisms and principles of learning in the brain. Sequential dependencies occur across domains of cognition -- perception, attention, categorization, decision making, language, choice -- and at multiple time scales.  We will examine both experimental papers and models that have been built to explain sequential dependencies.  And throughout the course, we will seek common underlying mechanisms and normative principles to explain sequential dependencies.

The course is open to any students who have some background in cognitive science or artificial intelligence.  Some background in proababilty and statistics will be helpful, but iis not essential as long as you are willing to learn.

In the style of graduate seminars, your primary responsibility for the course will be to read the series of papers before class and be prepared to come into class to discuss the paper (asking clarification questions, working through the math in the paper, relating the paper to other readings, critiquing the paper, presenting original ideas related to the paper).

For some of the readings, we'll ask you to write a one-page commentary on the paper, The commentary consists of approximately one page of comments, questions, or critiques of the assigned reading(s) for that class. This page will be due the day of class, and can include one or more of the following:

These commentaries are intended to promote careful thought about a paper before the session in which it is discussed. The point is not to give you more busy work, but rather to encourage you to jot down notes and questions as you read the papers. They will not be accepted after the class in which the paper is discussed.

You are required to present  a share of the papers during the course of the semester.  The presentation is meant to be a summary of the paper and its main ideas.  Ideally, two class members will collaborate to do each presentation, allowing you to work through the papers together.  We expect grad students to do twice the presentations that undergrads do.  We guess that grad students will do 2 or 3 presentations and undergrads will do 1.  

Presentations may be of the main article for the class (which everyone is required to read), or a supplementary article that we recommend.  By assigning students to present the supplementary articles, we can cover a lot more material without asking everyone to read every paper.

Grades will be based roughly on the following:  oral presentations 20%, class discussions 20%, written commentary on papers 60%.

All papers for the course can be found here or click on the individual readings.

Date	Activity	Reading	Supplementary Reading(s)	presenter
1/15	Course introduction
1/17	Vision and eye movements	Fecteau & Munoz (2003) Exploring the consequences of the previous trial		Mozer
1/22		Sharma et al. (2003) V1 neurons signal acquisition of an internal representation of stimulus location	Maloney, L. T., Dal Martello, M. F., Sahm, C., & Spillmann, L. (2005). Past trials influence perception of ambiguous motion quartets through pattern completion	Jones
1/24	Visual attention	Kristjansson (2006) Rapid learning in attention shifts:  A review	Kristjansson et al. (2006). Neural basis for priming of pop-out during visual search revealed with fMRI.	Mozer
1/29		Mozer, Shettel, & Vecera (2006).  Control of visual attention: A rational account COMMENTARY IS NOT REQUIRED	Mozer & Baldwin (2008).  Experience-guided search: A theory of attentional control	Mozer
1/31	Distinct mechanisms underlying sequential effects	Cho, Nystrom, Brown, Jones, Braver, Holmes, & Cohen (2002). Mechanisms underlying dependencies of performance on stimulus history in a two alternative forced choice task		Jones
2/5		Jentzsch & Sommer (2002) Functional localization and mechanisms of sequential effects in serial reaction time COMMENTARY IS  REQUIRED	Notebaert & Soetens (2003) . The influence of irrelevant stimulus changes on stimulus and response repetition effects.	Jones
2/7	Motor control and response intiation	Mozer, Kinoshita, & Davis (2004) Control of response initiation: Mechanisms of adaptation to recent experience COMMENTARY IS NOT REQUIRED	Song & Nakayama (2007). Automatic adjustment of visuomotor readiness	Mozer; Jones (Song & Nakayama)
2/12		Dixon & Glover (2004). Action and memory.		Jones
2/14	Sequence perception	Gilovich, T., Vallone, R., & Tversky, A. (1985). The hot hand in basketball: On the misperception of random sequences	Gilden & Wilson (1995) Streaks in skilled performance Bar-Eli, Avugos, & Raab (2006).  Twenty years of “hot hand” research: Review and critique	Matt Wilder, Dan Knights (Gilovich); Holger Dick (Gilden)
2/19		Steyvers & Brown (2006). Optimal change detection	Johnson J, Tellis GJ (2005).  Blowing bubbles: Heuristics and biases in the run-up of stock prices	Adam Bates (Johnson & Tellis); Mozer (Steyvers & Brown)
2/21	Perceptual judgement and estimation	DeCarlo & Cross (1990) Sequential effects in magnitude estimation: Models and theory	Jesteadt, Luce, & Green (1977) Sequential effects in judgments of loudness	Jones
2/26		Treisman & Williams (1984) A theory of criterion setting with an application to sequential dependencies	Lockhead & King (1983) A memory model for sequential effects in scaling tasks	Mozer  & Adam Bates
2/28		Petrov & Anderson (2005) The dynamics of scaling COMMENTARY IS NOT REQUIRED, BUT TRY TO HAVE A LOOK AT THE PAPER (We decided that two heavy papers in one week was too much.)	Petrov, Dosher, & Lu (2005). The dynamics of perceptual learning: An incremental reweighting model Petrov, Dosher, & Lu (2006). Perceptual learning without feedback in non-stationary contexts: Data and model	Jones
3/4		Stewart, Brown, & Chater (2005). Absolute identification by relative judgement	Brown, Marley, & Lacouture (2007) Is absolute identification always relative? Stewart, N. (2007). Absolute identification is relative: A reply to Brown, Marley, and Lacouture	Laura Rassbach (Stewart, Brown & Chater, 2005)
3/6	Probability learning	Myers, JL (1976). Probability learning. In W. K. Estes (Ed.), Handbook of learning and cognitive processes: Vol. 3. Approaches to human learning and motivation (pp. 171-205). Hillsdale, NJ: Erlbaum.	Sutton & Barto (1998) Reinforcement learning, Section 2.6 Estes (1957) Theory of learning with constant, variable, or contingent probabilities of reinforcement Anderson (1960) Effects of first-order conditional probability in a two-choice learning situation	Matt Wilder, Dan Knights
3/11		SUMMARY DAY		Mozer & Jones
3/13		Gallistel et al.(2001)  The rat approximates an ideal detector of changes in rates of reward		Tae Kwon
3/18		Sugrue, Corrado, & Newsome (2004). Matching behavior and the representation of value in the parietal cortex. Science, 304, 1782-1787.		Adam Bates
3/20	Spacing of Practice		Landauer, T K (1986) How Much Do People Remember? Some Estimates of the Quantity of Learned Information in Long-term Memory, Cognitive Science 10, 477-493. Landauer, T. K. (1975) Memory Without Organization: Properties of a Model with Random Storage and Undirected Retrieval, Cognitive Psychology, 7, 495-531.	Tom Landauer
3/25	SPRING BREAK
3/27
4/1	Probability learning	Behrens et al. (2007)  Learning the value of information in an uncertain world	Supplementary material for article	Laura Rassbach
4/3	Foraging	Cuthill, Kacelnik, Krebs, Haccou, & Iwasa (1990) Starlings exploiting patches: The effect of recent experience on foraging decisions	Real, L. A. (1991). Animal choice behavior and the evolution of cognitive architecture Dukas & Real (1993) Effects of nectar variance on learning by bumble bees	Tres Spicher (Cuthill et al., 1990); Holger Dick (Real, 1991); Ron Le Bel (Dukas & Real, 1993)
4/8	Causality	Chapman (1991) Trial order affects cue interaction in coningency judgment	Lopez, Shanks, Almaraz, & Fernandez (1998) Effects of trial order on contingency judgments: a comparison of associative and probabilistic contrast accounts	Kelsey Anderson (Lopez et al.)
4/10		Matute, Vegas, & Marez (2002) Flexible use of recent information in causal and predictive judgments		Mark Lewis-Pranzen
4/15 (Mozer away)	Category Learning	Jones & Sieck (2003). Learning myopia: An adaptive recency effect in category learning	Stewart, Brown, & Chater (2002) Sequence effects in categorization of simple perceptual stimul	Jones
4/17 (Mozer away)		Jones, Love, & Maddox (2006) Recency effects as a window to generalization	Jones, Maddox, & Love (2006) Stimulus generalization in category learning	Jones (main reading); Hadjar Homaei (supplementary paper)
4/22		Sakamoto, Jones, & Love (2008) Putting the psychology back into psychological models: Mechanistic vs. rational approaches	Nosofsky, Kruschke, & McKinley (1992) Combining exemplar-based category representations and connectionist learning rules	Jones; Hadjar Homaei (Nosofsky et al.)
4/24	Memory	Anderson, J. R.  & Schooler, L. (1991) Reflections of the environment in memory, Psychological Science, 2, 396-408.	Anderson, Tweney, Rivardo, & Duncan (1997) Need probability affects retention: A direct demonstration	Tres Spicher (Anderson & Schooler, 1991); Braden Wright (Anderson et al., 1997)
4/29		Brown, Steyvers, & Hemmer (2007). Modeling experimentally induced strategy shifts	Kruschke, J. (2006). Locally Bayesian learning. Brown S, Steyvers M. (2005).  The dynamics of experimentally induced criterion shifts	Ron Le Bel (Brown et al., 2007); Hadjar Homaei (Kruschke, 2006)
5/1	Long-term dependencies	Gilden (2001) Cognitive emissions of 1/f noise	Sanborn & Griffiths. (2008). MCMC with people. Kello, Beltz, Holden, & van Orden (2007) The emergent coordination of cognitive function	Mark Lewis-Pranzen (Gilden, 2001); Adam Bates (Sanborn & Griffiths)

Overview

Stimulus and Response sequences -- alternation, priming of repetition, response priming

Task Difficulty

            Marios G. Philiastides, Roger Ratcliff, and Paul Sajda1. Neural representation of task difficulty and decision making during perceptual categorization:  A timing diagram.

Attention

Categorization

Memory

Judgement

Causal Learning

Probability & Reinforcement Learning

Conditioning

Language

Game theory

Vlaev & Chater (2006) Game relativity: How context influences strategic decision making
Jones & Zhang (2004) Rationality and bounded information in repeated games, with application to the iterated Prisoner’s Dilemma
Colman (1998). Rationality assumptions of game theory and the backward induction paradox. In: Rational models of cognition, ed. M. Oaksford & N. Chater. Oxford University Press. (BF311.R34 1998)

Long-term effects

Decision Making