Cleotilde Gonzalez

Instance-Based Learning: Integrating Sampling and Repeated Decisions from Experience.

In decisions from experience, there are 2 experimental paradigms: sampling and repeated-choice. In the sampling paradigm, participants sample between 2 options as many times as they want (i.e., the stopping point is variable), observe the outcome with no real consequences each time, and finally select 1 of the 2 options that cause them to earn or lose money. In the repeated-choice paradigm, participants select 1 of the 2 options for a fixed number of times and receive immediate outcome feedback that affects their earnings. These 2 experimental paradigms have been studied independently, and different cognitive processes have often been assumed to take place in each, as represented in widely diverse computational models. We demonstrate that behavior in these 2 paradigms relies upon common cognitive processes proposed by the instance-based learning theory (IBLT; Gonzalez, Lerch, & Lebiere, 2003) and that the stopping point is the only difference between the 2 paradigms. A single cognitive model based on IBLT (with an added stopping point rule in the sampling paradigm) captures human choices and predicts the sequence of choice selections across both paradigms. We integrate the paradigms through quantitative model comparison, where IBLT outperforms the best models created for each paradigm separately. We discuss the implications for the psychology of decision making.

Does animation in user interfaces improve decision making

This paper reports a laboratory experiment that investigated the relative effects of images, transitions, and interactivity styles used in animated interfaces in two decision making domains. Interfaces used either realistic or abstract images, smooth or abrupt transitions, and parallel or sequential interactivity. Results suggest that decision making performance is influenced by the task domain, the user experience, the image, transition, and interactivity styles used in animated interfaces. Subjects performed better with animated interfaces based on realistic rather than abstract images. Subjects were more accurate with smooth rather than abrupt animation. Subjects were more accurate and enjoyed more the animation with parallel rather than sequential interactivity. Implications on the design of animated interfaces for decision making are provided. To account for appropriateness and interactivity, animation in HCI can be defined as: a series of varying images presented dynamically according to user actions in ways that help the user to perceive a continuous change over time and develop a more appropriate mental model of the task [7]. A task is generally considered a meaningful unit of work performance, and interactivity is usually defined by the visible or motor actions the user performs on the interface. This paper summarizes previous empirical research on animation and reports a laboratory study conducted to evaluate the decision making effectiveness of different types of images, transitions, and interactivity styles used in animated interfaces. The paper concludes with implications and recommendations concerning the design of animated interfaces for supporting decision making.

Cyber Situation Awareness: Modeling Detection of Cyber Attacks With Instance-Based Learning Theory

Objective: To determine the effects of an adversary’s behavior on the defender’s accurate and timely detection of network threats. Background: Cyber attacks cause major work disruption. It is important to understand how a defender’s behavior (experience and tolerance to threats), as well as adversarial behavior (attack strategy), might impact the detection of threats. In this article, we use cognitive modeling to make predictions regarding these factors. Method: Different model types representing a defender, based on Instance-Based Learning Theory (IBLT), faced different adversarial behaviors. A defender’s model was defined by experience of threats: threat-prone (90% threats and 10% nonthreats) and nonthreat-prone (10% threats and 90% nonthreats); and different tolerance levels to threats: risk-averse (model declares a cyber attack after perceiving one threat out of eight total) and risk-seeking (model declares a cyber attack after perceiving seven threats out of eight total). Adversarial behavior is simulated by considering different attack strategies: patient (threats occur late) and impatient (threats occur early). Results: For an impatient strategy, risk-averse models with threat-prone experiences show improved detection compared with risk-seeking models with nonthreat-prone experiences; however, the same is not true for a patient strategy. Conclusions: Based upon model predictions, a defender’s prior threat experiences and his or her tolerance to threats are likely to predict detection accuracy; but considering the nature of adversarial behavior is also important. Application: Decision-support tools that consider the role of a defender’s experience and tolerance to threats along with the nature of adversarial behavior are likely to improve a defender’s overall threat detection.

How choice ecology influences search in decisions from experience

Research into human decision-making has often sidestepped the question of search despite its importance across a wide range of domains such as search for food, mates, allies, visual targets or information. Recently, research on decisions from experience has made progress in finding out how individual characteristics shape search for information. Surprisingly little is known, however, about how the properties of the choice ecology shape peoples search. To fill this void, we analyzed how two key ecological properties influence search effort: domain of choice (gains vs. losses) and experienced variance (variance vs. no variance). Many people search longer when facing the prospect of losses relative to gains. Moreover, most people search more in options in which they experience variance relative to options they experience as invariant. We conclude that two factors that have been identified as important determinants of choice also influence search of information.

Effects of cyber security knowledge on attack detection

We quantitatively evaluate the role of knowledge when detecting cyber-attacks.Knowledge supports the identification of the relevant cues for classifying events.Knowledge facilitates integration of cues when detecting malicious network events.Knowledge makes a decision maker more aware of the type of cyber-attack (context).Situated knowledge is crucial to correctly integrate events and detect a cyber-attack. Ensuring cyber security is a complex task that relies on domain knowledge and requires cognitive abilities to determine possible threats from large amounts of network data. This study investigates how knowledge in network operations and information security influence the detection of intrusions in a simple network. We developed a simplified Intrusion Detection System (IDS), which allows us to examine how individuals with or without knowledge in cyber security detect malicious events and declare an attack based on a sequence of network events. Our results indicate that more knowledge in cyber security facilitated the correct detection of malicious events and decreased the false classification of benign events as malicious. However, knowledge had less contribution when judging whether a sequence of events representing a cyber-attack. While knowledge of cyber security helps in the detection of malicious events, situated knowledge regarding a specific network at hand is needed to make accurate detection decisions. Responses from participants that have knowledge in cyber security indicated that they were able to distinguish between different types of cyber-attacks, whereas novice participants were not sensitive to the attack types. We explain how these findings relate to cognitive processes and we discuss their implications for improving cyber security.

A cognitive modeling account of simultaneous learning and fatigue effects

Current understanding of sources of fatigue and of how fatigue affects performance in prolonged cognitive tasks is limited. We have observed that participants improve in response time but decrease in accuracy after extended repetitive work in a data entry task. We attributed the increase in errors to accumulating fatigue and the reduction in response time to learning. The concurrent effects of fatigue and learning seem intuitively reasonable but have not been explained computationally. This paper presents a cognitive computational model of these effects. The model, developed using the ACT-R cognitive architecture (Anderson et al., 2004; Anderson & Lebiere, 1998), accounts for learning and fatigue effects through a time-dependent modification of architectural parameters. The model is tested against human data from two independent experiments. Best fit to human accuracy and total response time was found from a modulation of both cognitive and arousal processes. Implications for training and skill acquisition research are discussed.

A Loser Can Be a Winner: Comparison of Two Instance-based Learning Models in a Market Entry Competition

This paper presents a case of parsimony and generalization in model comparisons. We submitted two versions of the same cognitive model to the Market Entry Competition (MEC), which involved four-person and two-alternative (enter or stay out) games. Our model was designed according to the Instance-Based Learning Theory (IBLT). The two versions of the model assumed the same cognitive principles of decision making and learning in the MEC. The only difference between the two models was the assumption of homogeneity among the four participants: one model assumed homogeneous participants (IBL-same) while the other model assumed heterogeneous participants (IBL-different). The IBL-same model involved three free parameters in total while the IBL-different involved 12 free parameters, i.e., three free parameters for each of the four participants. The IBL-different model outperformed the IBL-same model in the competition, but after exposing the models to a more challenging generalization test (the Technion Prediction Tournament), the IBL-same model outperformed the IBL-different model. Thus, a loser can be a winner depending on the generalization conditions used to compare models. We describe the models and the process by which we reach these conclusions.

A generic dynamic control task for behavioral research and education

Recent research in behavioral sciences presents strong evidence of poor human understanding for dynamic systems. Computer-based dynamic control tasks have an important potential for helping behavioral scientists advance research that investigates reasons for poor understanding and for helping students understand how dynamic systems work. In this paper, we introduce a simulation called Dynamic Stocks and Flows (DSF) that portrays the basic building blocks of dynamic systems: an accumulation; an inflow and outflow determined by an environment; and an inflow and outflow determined by a decision maker. In DSF, decision makers control the accumulation to a goal level by making repeated inflow and outflow decisions. We provide details of an experiment conducted with DSF that highlight some problems people face in controlling a dynamic system with different kinds of environmental inflow and outflow functions. DSF is flexible enough to represent dynamic systems with continuous or discrete accumulations, and with real-time or event-driven decision-making. We suggest that these and other features in DSF make it a good research and educational tool.

Situation Awareness in Dynamic Decision Making: Effects of Practice and Working Memory

The goal of this research is (a) to investigate the effects of task practice on situation awareness (SA); (b) to investigate how cognitive ability – in particular, working memory – moderates this practice effect on SA; and (c) to investigate effects of the SA measurement procedure (covering or uncovering the display while queries are answered). Task practice and working memory influence SA. However, the dynamics of the relationship between working memory and SA over time are not well understood, particularly with regard to using different SA measures. This research reports an experiment in which different SA measurement methods were used while participants played a computer simulation over several days. SA was measured using two query methods: a covered method in which questions were asked while the display was blanked out, and an uncovered method in which questions were asked while the display was shown. A working memory measure was collected from participants. Results indicate that the relationship between working memory and SA diminishes with task practice and differs between a covered and uncovered method. We also found that SA improvement with practice depends on the way in which SA is measured. We discuss the implications of these findings for understanding SA development, SA measurement, and the relationship of working memory in this process. The results from this study have clear implications for systems design, for the design of learning aids, and for SA measurement.

Dissociation of S-R compatibility and Simon effects with mixed tasks and mappings.

Binary-choice reactions are typically faster when the stimulus location corresponds with that of the response than when it does not. This advantage of spatial correspondence is known as the stimulus-response compatibility (SRC) effect when the mapping of stimulus location, as the relevant stimulus dimension, is varied to be compatible or incompatible with response location. It is called the Simon effect when stimulus location is task-irrelevant. The SRC effect is eliminated when compatible and incompatible spatial mappings are mixed within a trial block, and the Simon effect is eliminated when the Simon task is mixed with the SRC task with incompatible spatial mapping. Eliminations of both types have been attributed to suppression of an automatic response-activation route. We tested predictions of this suppression hypothesis for conditions in which the SRC and Simon tasks were intermixed and the spatial mappings on the SRC trials could be compatible or incompatible. In Experiment 1, the two tasks were equally likely, as were compatible and incompatible spatial mappings on SRC trials; in Experiment 2, the SRC or Simon task was more frequent; and, in Experiment 3, the compatible or incompatible location mapping for the SRC task was more frequent. The SRC effect was absent overall in all experiments, whereas the Simon effect was robust to the manipulations and showed the characteristic decrease across the reaction time (RT) distribution. This dissociation of effects implies that the automatic response-activation route is not suppressed in mixed conditions and suggests that mixing influences the SRC and Simon effects by different means.