How to Cite
Abstract
Background: Several techniques allow the development of poorly predictive artificial behavior models, such as finite state machines (FSM) and the use of cognitive architectures based on the theory of mind for the construction of agents whose behavioral model lies on a system of productions. Objective: It was proposed to generate artificial behavioral models to determine the conditions under which they demonstrate poorly predictive behaviors. Methodology: The first stage consisted of choosing platforms and tools; Pogamut, UT2000, SOAR, and Java were chosen. In the second stage, the coupling interface between the cognition engine, the language, and the environment was developed. Finally, in the third stage, the behavioral models were tested. Results: In the FSM model, it was possible to contrast the states and decisions made by the agents when there are restrictions in the set of actions predefined in its logic. It was also possible to contrast SOAR productions in terms of the predictability of the agent's actions based on what is perceived in the environment. Conclusions: Finite state machines are an important component when one wants to inspect the reactive behavior of an agent that pursues a single goal. Reflexive agents rely on their logic for the immediate perception of their environment without regard to decisions they have made or states they have already undergone. SOAR programs adjust the feedback of their environments in certain cases.
References
Chakraborti, T., Kulkarni, A., Sreedharan, S., Smith,, D. E., & Kambhampati, S. (2019). Explicability? Legibility? Predictability? Transparency? Privacy? Security? The Emerging Landscape of Interpretable Agent Behavior. In J. Benton (Ed) Proceedings of the Twenty-Ninth International Conference on Automated Planning and Scheduling (Vol. 29) (ICAPS 2019) (pp. 86 - 96). Association for the Advancement of Artificial Intelligence.
Champandard, A. (2003). AI Game Development: Synthetic Creatures with Learning and Reactive Behaviors. New Riders Publishing.
Laird, J., Newell, A. & Rosenbloom, P. (1987). SOAR: An Architecture for General Intelligence. The Journal of Artificial Intelligence [AIJ], 33(1), 1-64. https://doi.org/10.1016/0004-3702(87)90050-6
Langley, P., Laird, J. & Rogers, S. (2008). Cognitive architectures: Research issues and challenges. (Elsevier, Ed.) Cognitive Systems Research, 10(2), 143-158. https://doi.org/10.1016/j.cogsys.2006.07.004
Mora, A., Castillo, P. A., García-Sanchez, P. & Merelo, J. J. (2015). Modelling a Human-Like Bot in a First Person Shooter Game. International Journal of Creative Interfaces and Computer Graphics, 6(1), 21-37. https://doi.org/10.4018/IJCICG.2015010102
Norvig, P. & Russell, S. (2010). A.I. A modern Approach. 3rd Ed. New Jersey: Pearson.
Petersen, S. E., & Sporns, O. (2015). Brain Networks and Cognitive Architectures. Neuron Perspective, 88(1), 207 – 219. https://dx.doi.org/10.1016/j.neuron.2015.09.027
Vélez B., J. I., Castillo O., L. F. & González, M. (2010). Implementación de un modelo de comportamiento reactivo para agentes en un entorno de videojuegos. Revista Vector, 5(1), 61 - 68. http://vector.ucaldas.edu.co/downloads/Vector5_7.pdf
This work is licensed under a Licencia Creative Commons Atribución-NoComercial-