Robert L. West

Simple games as dynamic, coupled systems: randomness and other emergent properties

From a game theory perspective the ability to generate random behaviors is critical. However, psychological studies have consistently found that individuals are poor at behaving randomly. In this paper we investigated the possibility that the randomness mechanism lies not within the individual players but in the interaction between the players. Provided that players are influenced by their opponents past behavior, their relationship may constitute a state of reciprocal causation [Cognitive Science 21 (1998) 461], in which each player simultaneously affects and is affected by the other player. The result of this would be a dynamic, coupled system. Using neural networks to represent the individual players in a game of paper, rock, and scissors, a model of this process was developed and shown to be capable of generating chaos-like behaviors as an emergent property. In addition, it was found that by manipulating the control parameters of the model, corresponding to the amount of working memory and the perceived values of different outcomes, that the game could be biased in favor of one player over the other, an outcome not predicted by game theory. Human data was collected and the results show that the model accurately describes human behavior. The results and the model are discussed in light of recent theoretical advances in dynamic systems theory and cognition.

On the reliability of usability testing

Six professional usability testing teams conducted a usability test on an early prototype of a dialog box. Altogether, they identified 36 usability problems. No problem was detected by every team, 2 were found by five teams, 4 by four teams, 7 by three teams, 7 by two teams, and 18 problems were identified by one team only. There was more agreement among teams in this study compared to a previous study [1] and there was more agreement among the teams on severe vs. minor problems. Implications for the cooperation between usability testers and their clients are discussed.

Deconstructing and reconstructing ACT-R: Exploring the architectural space

Evaluating variations in the structure of computational models of cognition is as important as evaluating variations in the numerical parameters of such models. However, computational models tend not to be organized in such a way as to directly support such research. To address this need, we have taken the well-known cognitive architecture ACT-R, reduced it to its fundamental components, and reconstructed it. Our new system, Python ACT-R, facilitates exploration of the space of possible models and architectures based on the core ACT-R theory. The result has enabled us to examine the possibility of using basic ACT-R components such as the declarative memory system in new ways; for example, as the basis for a new visual attention system. Python ACT-R allows the same model definition syntax to be used to define both ACT-R models and new ACT-R components, as well as making explicit the processes specified by the ACT-R theory.

Positive and negative congruency effects in masked priming: A neuro-computational model based on representation, attention, and conflict

Studies on masked and unmasked priming have long shown reliable positive effects of the congruent prime on target processing. Paradoxically, a negative effect has also been found, showing faster and more accurate responses in the incongruent compared to the congruent trials. Positive effects have been found with a short time between the prime and the target, while negative effects have been found with a long time between the prime and the target. This has been modeled by assuming that the prime initiates a motor self-inhibitory process that causes these effects (Bowman, H., Schlaghecken, F., Eimer, M., 2006. A neural network model of inhibitory processes and cognitive control. Vis. Cogn. 13, 401-480). We have developed an alternative explanation based on attentional neuro-modulation. In this paper we show that attentional neuro-modulation can be used to model a wide range of findings in this area.

Using GOMS for Modeling Routine Tasks Within Complex Sociotechnical Systems: Connecting Macrocognitive Models to Microcognition

Cognitive modeling has not yet played much of a role in the study of sociotechnical systems. Arguably, this is because most cognitive modeling systems were originally created to model microcognitive results, not the types of macrocognitive behaviors that drive sociotechnical systems (Klein et al., 2003). However, this does not mean that cognitive modeling systems cannot be adapted to deal with macrocognitive activities in ways that are relevant to cognitive engineering. Previous research using GOMS in sociotechnical systems indicated that GOMS is problematic to use when interruptions and task switching are common; therefore, we added new theoretical structures to GOMS to deal with these issues. We tested the system by constructing a model of routine network maintenance and installation at a large telecommunications company. We then compared the model predictions with observations of the work. The results showed that the model results were useful in guiding the research and organizing the findings.

Testing for Equivalence: A Methodology for Computational Cognitive Modelling

Testing for Equivalence: A Methodology for Computational Cognitive Modelling The equivalence test (Stewart and West, 2007; Stewart, 2007) is a statistical measure for evaluating the similarity between a model and the system being modelled. It is designed to avoid over-fitting and to generate an easily interpretable summary of the quality of a model. We apply the equivalence test to two tasks: Repeated Binary Choice (Erev et al., 2010) and Dynamic Stocks and Flows (Gonzalez and Dutt, 2007). In the first case, we find a broad range of statistically equivalent models (and win a prediction competition) while identifying particular aspects of the task that are not yet adequately captured. In the second case, we re-evaluate results from the Dynamic Stocks and Flows challenge, demonstrating how our method emphasizes the breadth of coverage of a model and how it can be used for comparing different models. We argue that the explanatory power of models hinges on numerical similarity to empirical data over a broad set of measures.

A holographic associative memory recommender system

We describe a recommender system based on dynamically structured holographic memory (DSHM), a cognitive model of associative memory that uses holographic reduced representations as the basis for its encoding of object associations. We compare this recommender to a conventional user-based collaborative filtering algorithm on three datasets: MovieLens, and two bibliographic datasets such as those typically found in a digital library. Off-line experiments show that the holographic recommender is competitive in accuracy for predicting movie preferences and more accurate than collaborative filtering on very sparse data sets. However, DSHM requires significant amounts of computational resources which may require a distributed implementation for it to be practical as a recommender for large data sets.

Cyberpsychology: A Human-Interaction Perspective Based on Cognitive Modeling

This paper argues for the relevance of cognitive modeling and cognitive architectures to cyberpsychology. From a human-computer interaction point of view, cognitive modeling can have benefits both for theory and model building, and for the design and evaluation of sociotechnical systems usability. Cognitive modeling research applied to human-computer interaction has two complimentary objectives: (1) to develop theories and computational models of human interactive behavior with information and collaborative technologies, and (2) to use the computational models as building blocks for the design, implementation, and evaluation of interactive technologies. From the perspective of building theories and models, cognitive modeling offers the possibility to anchor cyberpsychology theories and models into cognitive architectures. From the perspective of the design and evaluation of socio-technical systems, cognitive models can provide the basis for simulated users, which can play an important role in usability testing. As an example of application of cognitive modeling to technology design, the paper presents a simulation of interactive behavior with five different adaptive menu algorithms: random, fixed, stacked, frequency based, and activation based. Results of the simulation indicate that fixed menu positions seem to offer the best support for classification like tasks such as filing e-mails. This research is part of the Human-Computer Interaction, and the Broadband Visual Communication research programs at the National Research Council of Canada, in collaboration with the Carleton Cognitive Modeling Lab at Carleton University.

Modeling SGOMS in ACT-R: Linking Macro- and Microcognition

West and Nagy (2007) first addressed the issue of using cognitive architectures for modeling macrocognition by arguing that this can be viewed as a special case of macrocognition, in which there is a deliberate attempt to connect macrocognition to microcognition through the use of the architecture. West and Nagy also developed and tested a method for applying GOMS, (a modeling system based on Goals, Operators, Methods, and Selection Rules created by Card, Moran, and Newell, 1983) to model macrocognition. That system was called Sociotechnical GOMS (SGOMS). In this paper we further discuss the relationship between cognitive modeling and macrocognition and describe our work on implementing the SGOMS system in the ACT-R cognitive architecture.

A GOMS model of virtual sociotechnical systems: using video games to build cognitive models

Motivation -- The present paper extends the use of GOMS models, described by Kieras (Kieras, 2007) as models of the knowledge necessary for an agent to perform a task, to complex sociotechnical processes involving multiple agents in strategic activities situated in a virtual environment. Research approach -- The experiment consists of the SGOMS model, based on task analysis, and a statistical analysis to evaluate the accuracy of the SGOMS model for the description and prediction of the data. Findings/Design -- A SGOMS model featuring task interruptions, order violations and planning units representing the decision-making process is a good match with the experimental data. Take away message -- The GOMS model can be modified to account for complex sociotechnical interactions within low-fidelity synthetic environments.