Citation Context

...fs on each trial are dependent on the proportion of selections made to each option over a 20-trial moving window. Thus, this proportion of responses defines the current state of the task environment (=-=Gureckis & Love, 2009-=-). If the next response changes the relative proportion of the previous 20 responses, then the state changes. In the rising optimum, there are 21 possible states. 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 Rising ...

Citation Context

... to maximize gains, making use of a reward signal that provides information about the “goodness” of actions. This framework has been used to model human decision-making behavior (Fu & Anderson, 2006; =-=Gureckis & Love, 2009-=-) as well as firing patterns in dopaminergic neurons in primates (Schultz, Dayan, & Montague, 1997). Our RL model demonstrates that weighting of a fictive (i.e., forgone) reward signal for the action ...

Citation Context

...3.3%, Caucasian: 46.5%, Native American: <0.1%, Others: 1.5%. The ages of participants in this pool ranged from 17 to 55 (M ! 19.08, SD ! 1.76). Materials and procedure. Participants were administered the BIS-11 questionnaire (Patton et al., 1995) that consists of 30 statements, such as ‘‘I do things without thinking’’ and ‘‘I am more interested in the present than the future’’ with which participants stated their level of agreement on a 4-point scale. Higher summed scores indicate higher levels of impulsivity. Following the questionnaire, participants played the ‘‘Farming on Mars’’ game (see Gureckis & Love, 2009), an adaptation of Herrnstein et al.’s (1993) choice task. Participants read a story about National Aeronautics and Space Administration (NASA) scientists on Mars attempting to extract oxygen from its atmosphere in order to create breathable air for use in a human colony. They were informed that, as members of the project, their job was to extract as much oxygen as possible from the atmosphere. To do this, they needed to repeatedly choose between two ‘‘oxygen-extraction robots’’ with different properties. Beyond this information, participants were only told that the specific oxygen-extracting ...

Citation Context

...nd feedback can scale up to explain effects known as melioration of performance, where maximization of short-term gains requires different action sequences than maximization of long-term gains (e.g., =-=Gureckis & Love, 2009-=-; Herrnstein, 1990; Neth, Sims, & Gray, 2006; Shanks, Tunney, & McCarthy, 2002). The magnitude of rewards can vary more easily between alternative settings and will depend on the objective function at...

Citation Context

...gy that maximizes the highest local reward (e.g., the reward value after a single visit) rather than finding the strategy that maximizes the cumulative reward to be had at the end of the trial (e.g., =-=Gureckis & Love, 2009-=-; Neth, Sims, & Gray, 2006). Participants Method Twenty Master’s students (seven female) from University College London participated on a voluntary basis. Participants were aged between 22 and 37 year...

Citation Context

...of their simplicity and unexamined intuitions about the “randomness” of exploratory behavior, reflexive approaches are commonly adopted to model human behavior (Daw et al., 2006; Worthy et al., 2007; =-=Gureckis and Love, 2009-=-; Pearson et al., 2009; Jepma and Nieuwenhuis, 2011). Reflexive approaches are also prominent in the design of artificial agents (Sutton and Barto, 1998). THE LEAPFROG VARIANT OF THE CLASSIC BANDIT TA...

Citation Context

... Experiment 2, participants performed one of two dynamicsdecision-making tasks in which reward values were dependentson the sequence of previous choices (Bogacz, McClure, Li,sCohen, & Montague, 2007; =-=Gureckis & Love, 2009-=-; Otto,sGureckis, Markman, & Love, 2009). Figure 3 shows thesreward structures for the tasks. In each task, there were twosoptions: a decreasing option and an increasing option. Thesdecreasing option ...

Citation Context

... Experiment 2, participants performed one of two dynamicsdecision-making tasks in which reward values were dependentson the sequence of previous choices (Bogacz, McClure, Li,sCohen, & Montague, 2007; =-=Gureckis & Love, 2009-=-; Otto,sGureckis, Markman, & Love, 2009). Figure 3 shows thesreward structures for the tasks. In each task, there were twosoptions: a decreasing option and an increasing option. Thesdecreasing option ...

Citation Context

...imicking opponent and probability maximizing when encountering an indifferent opponent. Learning to choose optimally in our choice paradigm requires a number of cognitive processes also vital for decisions under uncertainty and competition in many real-world situations. These include the exploration of choice profitability and learning about the motives and choice strategies of competing agents. Computational models of reinforcement learning have been shown to successfully and parsimoniously describe such cognitive mechanisms in related tasks (e.g., Busemeyer & Stout, 2002; Erev & Roth, 1998; Gureckis & Love, 2009; Rieskamp & Otto, 2006) and provide an attractive approach to illuminating the nature of learning mechanisms adopted by decision makers in our paradigm. Following the presentation of the behavioral data, we therefore discuss the applicability of a variety of computational models of reinforcement learning that differ with regard to the importance they place on (solely) maximizing profit and out-smarting opponents.2. Experiment 1 2.1. Method 2.1.1. Participants Fifty (35 female) undergraduate students from the University of New South Wales with a mean age of 18.92 years (SD = 1.19 years) partic...

Citation Context

...ers, &Stout,2008; Ahn, Krawitz, Kim, Busemeyer, & Brown, 2011). The EV, PVL, and other RL models have been a dominant class of models used to characterize decisionmaking behavior in numerous studies (=-=Gureckis & Love, 2009-=-a, b; Worthy, Maddox, & Markman, 2007). The basic assumptions underpinning these RL models is that the outcomes of past decisions are integrated to determine expected reward values for each option, an...