Reinforcement-based decision making in corticostriatal circuits: mutual constraints by neurocomputational and diffusion models

In this letter, we examine the computational mechanisms of reinforce-ment-based decision making. We bridge the gap across multiple levels of analysis, from neural models of corticostriatal circuits-the basal ganglia (BG) model (Frank, 2005 , 2006 ) to simpler but mathematically tractable diffusion models of two-choice decision making. Specifically, we generated simulated data from the BG model and fit the diffusion model (Ratcliff, 1978 ) to it. The standard diffusion model fits underestimated response times under conditions of high response and reinforcement conflict. Follow-up fits showed good fits to the data both by increasing nondecision time and by raising decision thresholds as a function of conflict and by allowing this threshold to collapse with time. This profile captures the role and dynamics of the subthalamic nucleus in BG circuitry, and as such, parametric modulations of projection strengths from this nucleus were associated with parametric increases in decision boundary and its modulation by conflict. We then present data from a human reinforcement learning experiment involving decisions with low- and high-reinforcement conflict. Again, the standard model failed to fit the data, but we found that two variants similar to those that fit the BG model data fit the experimental data, thereby providing a convergence of theoretical accounts of complex interactive decision-making mechanisms consistent with available data. This work also demonstrates how to make modest modifications to diffusion models to summarize core computations of the BG model. The result is a better fit and understanding of reinforcement-based choice data than that which would have occurred with either model alone.     

Do websites influence the nature of voting intentions? The case of two national elections in Greece

This paper aims to contribute to the growing body of research on online political marketing by investigating the use of websites as a marketing tool in the 2007 and 2009 general elections in Greece. The two main research objectives are firstly to establish the key factors affecting voters’ trust when it comes to using politicians’ websites and secondly to examine whether and to what extent the influence of trust-building factors changes over-time, as the online environment matures. The key findings of this research are that users tend to visit the websites of favoured candidates, which reinforces their positive perceptions of them, while as far as the websites themselves and their content and the services provided are concerned there is still room for improvement.     

The effect of action video game experience on task-switching

There is now a substantial body of work demonstrating that action video game experience results in enhancements in a wide variety of perceptual skills. More recently, several groups have also demonstrated improvements in abilities that are more cognitive in nature, in particular, the ability to efficiently switch between tasks. In a series of four experiments, we add to this body of work, demonstrating that the action video game player advantage is not exclusively due to an ability to map manual responses onto arbitrary buttons, but rather generalizes to vocal responses, is not restricted to tasks that are perceptual in nature (e.g. respond to a physical dimension of the stimulus such as its color), but generalizes to more cognitive tasks (e.g. is a number odd or even), and is present whether the switch requires a goal-switch or only a motor switch. Finally, a training study establishes that the relationship between the reduction in switch cost and action game playing is causal.     

Bootstrapping Boosted Random Ferns for discriminative and efficient object classification

In this paper we show that the performance of binary classifiers based on Boosted Random Ferns can be significantly improved by appropriately bootstrapping the training step. This results in a classifier which is both highly discriminative and computationally efficient and is particularly suitable when only small sets of training images are available.During the learning process, a small set of labeled images is used to train the boosting binary classifier. The classifier is then evaluated over the training set and warped versions of the classified and misclassified patches are progressively added into the positive and negative sample sets for a new re-training step. In this paper we thoroughly study the conditions under which this bootstrapping scheme improves the detection rates. In particular we assess the quality of detection both as a function of the number of bootstrapping iterations and the size of the training set. We compare our algorithm against state-of-the-art approaches for several databases including faces, cars, motorbikes and horses, and show remarkable improvements in detection rates with just a few bootstrapping steps.     

Maximum margin Bayesian network classifiers

We present a maximum margin parameter learning algorithm for Bayesian network classifiers using a conjugate gradient (CG) method for optimization. In contrast to previous approaches, we maintain the normalization constraints on the parameters of the Bayesian network during optimization, i.e., the probabilistic interpretation of the model is not lost. This enables us to handle missing features in discriminatively optimized Bayesian networks. In experiments, we compare the classification performance of maximum margin parameter learning to conditional likelihood and maximum likelihood learning approaches. Discriminative parameter learning significantly outperforms generative maximum likelihood estimation for naive Bayes and tree augmented naive Bayes structures on all considered data sets. Furthermore, maximizing the margin dominates the conditional likelihood approach in terms of classification performance in most cases. We provide results for a recently proposed maximum margin optimization approach based on convex relaxation. While the classification results are highly similar, our CG-based optimization is computationally up to orders of magnitude faster. Margin-optimized Bayesian network classifiers achieve classification performance comparable to support vector machines (SVMs) using fewer parameters. Moreover, we show that unanticipated missing feature values during classification can be easily processed by discriminatively optimized Bayesian network classifiers, a case where discriminative classifiers usually require mechanisms to complete unknown feature values in the data first.     

Plate tectonic reconstructions with continuously closing plates

We present a new algorithm for modeling a self-consistent set of global plate polygons. Each plate polygon is composed of a finite list of plate margins, all with different Euler poles. We introduce a "continuously closed plate" (CCP), such that, as each margin moves independently, the plate polygon remains closed geometrically as a function of time. This method solves emerging needs in computational geodynamics to combine kinematic with dynamic models. Because they have polygons that are too widely spaced in time and have inconsistent motions between margins and plates, traditional global plate tectonic reconstructions have become inadequate for geodynamics. The CCP algorithm has been incorporated into the GPlates open-source paleogeographic system. The algorithm is a set of procedures and data structures that operate on collections of reconstructed geometric data to form closed plate polygons; the main data structures used for each plate polygon are based on a nested hierarchy of topological elements. Reconstructions with CCPs can be created, edited, visualized, and exported with GPlates. The native storage of the dynamic reconstructions is the GPlates Markup Language, GPML, which uses an XML-based file format called GML. We demonstrate the utility of the CCP method by creating a global reconstruction with continuously closing plates from 140 Ma to the present using data from existing, traditional reconstructions.     

SAT Solving for Termination Proofs with Recursive Path Orders and Dependency Pairs

This paper introduces a propositional encoding for recursive path orders (RPO), in connection with dependency pairs. Hence, we capture in a uniform setting all common instances of RPO, i.e., lexicographic path orders (LPO), multiset path orders (MPO), and lexicographic path orders with status (LPOS). This facilitates the application of SAT solvers for termination analysis of term rewrite systems (TRSs). We address four main inter-related issues and show how to encode them as satisfiability problems of propositional formulas that can be efficiently handled by SAT solving: (A) the lexicographic comparison w.r.t. a permutation of the arguments; (B) the multiset extension of a base order; (C) the combined search for a path order together with an argument filter to orient a set of inequalities; and (D) how the choice of the argument filter influences the set of inequalities that have to be oriented (so-called usable rules). We have implemented our contributions in the termination prover AProVE. Extensive experiments show that by our encoding and the application of SAT solvers one obtains speedups in orders of magnitude as well as increased termination proving power.