Cognitive Modeling Cognitive models are formalized mathematical or computational implementations of cognitive theories. Unlike statistical models, the parameters and general structure of the model are meant to correspond to mechanisms and processes important in human and animal cognition. These mechanisms may be structural and/or functional in nature. Animal Cognition Currently, we use a cognitive model within our ELAM framework to model fish movements. Sensory stimuli differentially activate four different response behaviors. Activation and decay functions are based on neural activation functions. The response behavior with the highest current activation is executed. Unlike a purely reactive system, behaviors may persist when stimuli no longer support the activation of that particular behavior. The ELAM model has been quite successful at predicting fish passage. The new flume will allow us to test the model more rigorously as well as to collect information to model more varied fish behaviors. Goodwin, R. A., Pandey, V., Kiker, G. A., and Kim, J. B. (2007). “Spatially-explicit population models with complex decisions: fish, cattle, and decision analysis.” In: Environmental Security in Harbors and Coastal Areas, NATO Security through Science Series, Springer, Netherlands, pp. 293-306. Human Cognition © iStockphoto.com / Vasiliy Yakobchuk Human beings are still the best example of a general purpose intelligence we know of. Nonetheless, our cognitive capabilities are very limited. Understanding such limitations has at least two direct uses. The first is to better understand how we operate in our own environments. Better understanding of the human element will help us to create new tools and processes that better integrate with humans within the system. Secondly, understanding how the human brain has evolved to trade off logical precision for speed and generality will help us design and implement the next generation of artificially intelligent machines. Cognitive Aspects of Decision Making Workshop (2008) Cognitive Decision Making Rational decision making has been the basis for many fields of study including economics, game theory, and artificial intelligence. Although such fields use rational normative models of human decision making, it has been understood for decades that such models are not accurate at describing how real humans make decisions. We are currently using dynamic evidence accumulation models of human decision making as the basis for developing a cognitive architecture focused on situated cognition. It is a non-trivial task to extend models developed to explain behaviors seen in the lab to explain behaviors in the real world. The real world does not control for variables not applicable to a specific point of study. Initial progress was made by theoretically extending and experimentally verifying how decision-making interacts with other cognitive elements such as motivation, learning, and planning. Coordinated Behaviors Social organisms are known to behave in coordinated fashion such that organized group behaviors cannot be derived from detailed understanding of the individual's behavior. Such group behaviors are known as emergent behaviors. Emergent systems can be studied using macro-level models that focus on the group itself as a unit of study. Additionally, micro-level models can produce emergent patterns by simulating the movements and interactions of individuals. Most of such simulations focus on how relatively simple relations and behavioral rules can yield observed group-level patterns. Unfortunately, the behaviors of people (and other social animals) are rarely accurately described using such simple rule sets. Modern computational resources allow us to apply complex cognitive models to behavioral scenarios involving multiple cognitive agents. Unlike others who are specifically researching group dynamics, we are interested in studying individuals in a group.