Cognitive Science and Artificial Intelligence

An introduction to the fundamental theories and issues on cognitive science, focusing on the organization of the human mind and other cognitive systems. This is an interdisciplinary class that brings together the fields of philosophy, psychology, mathematics and computer science/artificial intelligence. This class will introduce the fundamental concepts of cognitive science, decision making models, reinforcement learning and neural networks, ones that will have profound impact on every part of our society and give students a new way of thinking about problem solving and modeling. Hands-on exercise using MATLAB will help student understand the idea better and have a chance to utilize them.

Topics Learned

• Cognitive Science 
• Artificial Intelligence 
• Introduction to MATLAB 
• Reinforcement Learning 
• Evaluation Feedback 
• The Reinforcement Learning Problem 
• Dynamic Programming 
• Monte Carlo Methods 
• Temporal-Difference Learning 
• Generalization, Approximation and Unified View of Reinforcement Learning 
• Neural Networks: Perceptron, Back-Propagation, Recurrent Network 
• Adaptive Resonance Theory (ART1) Model

Project Abstract (paper available on request)

• A model was created to simulate an interaction with a repeated two choice decision-making task presented in a simple gambling situation. In the gambling task, two decks of cards of different monetary value are presented to a player. The goal is to maximize the total earnings over the entire experiment (1000 plays). Unknown to the agent is that a contingency is set up between the recent selection of card decks and the money received on the next trial. One deck of cards has a higher long-term reward while the other has a higher short-term reward. Specifically, the instant reward for the next trial is based on the history of the past four trials, that is, the number of choices made on the long-term deck. The short-term deck generates more reward for each individual trial, but over a period of four trials, repeatedly selecting the long-term goal will exceed that from repeatedly selecting the short-term goal. Obviously the optimal strategy is to select the long-term goal. As in Gureckis and Love [2009], we expect that small amounts of noise will make the agent more likely to converge to the optimal (that is, long-term deck selection) strategy, whereas large amounts of noise are likely to lead to instability which makes convergence to any strategy less optimal. The reinforcement learning model that was created explored ten levels of sigma while varying though seven levels of epsilon and found that more noise helped the model to optimize the long term goal.

Readings

• Thagard, P. (2005). Mind: Introduction to Cognitive Science, 2nd edition, The MIT Press, Cambridge, Massachusetts. 
• Sutton, R. S. & Barto, A. G. (1998). Reinforcement Learning: An Introduction, The MIT Press, Cambridge, Massachusetts.
• My Book notes for Reinforcement Learning: An Introduction
• Gurney, K. (1997). An Introduction to Neural Networks, CRC Press.
• Gureckis, T. M., & Love, B. C. (2009). Learning in noise: Dynamic decision-making in a variable environment. Journal of Mathematical Psychology, 53, 180-193

Research Design and Analysis II

This course is the advanced program in experi- mental design and analysis. The focus is the design, planning, and considerations involved in complex, multivariate experiments. Major areas of examination will include factorial designs, nested variables, linear models, multiple regression, measures of covariance, and Latin square designs. Considerations in selecting the appropriate experimental design is the focus of this course. Examination of appropriate statistical techniques is integrated with the theoretical and practical con- cepts of experimental design.

Topics Learned

• Analysis of Variance
• One-Way ANOVA
• A-priori Testing: Linear and Orthogonal Contrasts
• Post-hoc Tests
• Trend Analysis
• Factorial ANOVA
• 2-way ANOVA
• 3-way ANOVA
• Simple Effects
• Simple Comparisons
• Repeated Measures ANOVA
• 1-way Designs
• Mixed Designs
• ANCOVA
• Multivariate Analysis of Variance
• 1-way MANOVA
• Multiple Regression
• Introduction to Ordinary Least Squares
• Hierarchical Models
• Suppression
• Predictor Importance
• Model Design
• Partial and Semi-partial Correlation
• Diagnostics

Human Factors in Systems

Survey of human factors literature. Introduction to topics including human capabilities and human interfaces with human-machine systems, workload, anthropometrics, perception, workspace design, visual momentum. The course will study human limitations in the light of human engineering, human reliability, stress, and human physiology. The course will discuss human behavior as it relates to the aviator’s adaptation to flight, air traffic, and maintenance environments.

Topics Learned

• Research Methods
• Design and Evaluation
• Human Information Processing
• Displays
• Motor Skills and Controls
• Controls and Hand Tools
• Anthropometry and Workspace Design
• Situation Awareness and Cockpit Resource Management

Project Abstract (paper available on request)

• For several years, Florida’s Blood Centers have been utilizing a mobile blood donation center known as the Big Red Bus to solicit blood donations for local Florida hospitals. By traveling seven days a week to various workplaces, schools, places of worship, and shopping centers in the area, the Big Red Bus has been highly successful collecting donations because of the convenience it provides to the community. To accommodate the larger quantities of donors the Big Red Bus experiences at some locations, the bus has recently been redesigned. However, as with most new designs, there have been some design constraints which prevent the operation from running as efficiently and safely as would be ideal. Our research team observed the current donation process on the Big Red Bus and interviewed several of its employees, known as Donor Service Specialists (DSSs), to examine if there were any areas that needed to be redesigned to make the bus more user-friendly, both for the donor and the DSS. Our team discovered eight specific human factors and ergonomics-related issues aboard the bus and this paper describes those issues and their proposed solutions.

Readings

• Sanders, M. S., & McCormick, E. J. (1993). Human factors in engineering design (7th edition). McGraw-Hill: NY. ISBN: 0-07-054901-X. 
• Vincente, K. (2003). The Human Factor. Routledge: New York.
• Wickens, C.D., Lee, J. D., Liu, Y., Gordon Becker, S. E. (2004). An Introduction to Human Factors Engineering (2nd edition). Prentice Hall: Upper Saddle River, NJ. ISBN: 0-13-183736-2.
• Norman, D. A., (2002). The Design of Everyday Things. Basic Books. ISBN: 0-465-06710-7.
• Endsley, M. R. (1995a). Towards a theory of situational awareness in dynamic environments. Human Factors, 37, 32-64. 
• Endsley, M. R. (1995b). Measurement of situation awareness in dynamic systems. Human Factors, 37, 65-84. 
• Klein, G. (1993). A recognition primed decision (RPD) model of rapid decision making. In G. Klein, J. Orasanu, R. Calderwood, & C. E. Zsambok (eds.) Decision making in action: Models and methods (pp 138-147). Norwood, NJ: Ablex. 
• Lipshitz, R., Klein, G., Orasanu, J., & Salas, E. (2001). Focus article: Taking stock of naturalistic decision making. Journal of Behavioral Decision Making, 14, 331-352.

Human and Organizational Factors in Technological Systems

Theoretical paradigms in human-computer interaction and their application to interface design; advanced interface technologies such as multimodel input/output, hypertext, and knowledge- based systems. This class addresses the advanced topics on human and organizational factors by using a discrete event simulation (DES) approach. In this course, students will learn advanced techniques in simulation modeling and analysis using ARENA simulation software. Students will learn fundamental concepts/theory involved in discrete event simulation, including simulation structure and logic, simulation languages (ARENA), advanced HCI with simulation (Visual Basic Interface with ARENA), statistical analysis of the results and application to system situations.

Topics Learned

• Simulation Concepts
• ARENA
• Modeling Advanced Operations
• Statistical Analysis of Output
• Steady-state Statistics
• Entity Transfer
• Additional Modeling Issues
• ARENA Integration and Customization
• Conducting Simulation Studies

Project Abstract (paper available on request)

• The recommendations presented in this preliminary report to Sue’s Markets have been based on the optimized results of a customized Rockwell Arena 10 simulation developed from requirements provided by the stakeholder and based on end-user preferences. It was determined that an optimal part-time staff schedule for Sue’s Markets would employ nine cashiers and one bagger from the hours of 2:00 pm until 7:00 pm with seven cashiers and two baggers for the hours between 7:00 pm to 10:00 pm. This schedule should also include four dedicated “express” lanes that allow a customer to have no more than 131 purchase items. Following his schedule will yield minimal waiting time for customers at a minimal part-time staffing cost of $539.00 per day.

Readings

• Kelton, David et al (2007) Simulation with Arena, 4th edition, McGraw-Hill, 2007

Research Design and Analysis I

Foundation and procedures of research techniques, tools, and methods. Course reviews the principal concepts of research design and evaluation. The application of experimental, case-study, survey, and non-experimental techniques are explored. Identification, isolation, and treatment of dependent and independent variables covered. Existing published research or data used to highlight principles.

This course is designed to assist the student in the acquisition of basic research principles. This is achieved by exploring basic research methods, concepts, critically examining published research works, and active participation in research design exercises.

Performance Objectives:

• Basic research design models.
• Key design concepts.
• Contexts in which to use various research design models.
• Ability to critically think about experimental literature.
• Make formal presentation addressing a relevant psychological or human factors topic using appropriate terminology and support from the experimental literature.

Topics Learned

• Introduction/Explaining Behavior
• Developing  Ideas/Importance of Theory
• Choosing a Research Design
• Researching References
• Making Systematic Observations
• Selecting Participants
• Using Between and Within-Subject Designs
• APA Formatting
• Writing a Research/Thesis Proposal
• Preparing Research Presentations

Readings

• Association, A. P. Publication Manual of the American Psychological Association, Fifth Edition. American Psychological Association (APA).
• Bordens, K. and Abbott, B. B. (2007). Research Design and Methods: A Process Approach. McGraw-Hill Humanities/Social Sciences/Languages, New York, NY, USA, 7 edition.
• My Book notes for Research Design and Methods: A Process Approach

Human-Computer Interaction

This course stresses the importance of good interfaces and the relationship of user interface design to human-computer interaction. Other topics include interface quality and methods of evaluation interface design examples; dimensions of interface variability; dialogue genre; dialogue tools and techniques; user-centered design and task analysis; prototyping and the iterative design cycle; user interface implementation; prototyping tools and environments; I/O devices; basic computer graphics; color and sound.

Systems Engineering II

Studies of the value of prototyping in the application of design, build, and test processes. In-depth focus on the innovation of conceptual designs in short time-cycle engineering.