Qualitative reachability in stochastic BPA games

We consider a class of infinite-state stochastic games generated by stateless pushdown automata (or, equivalently, 1-exit recursive state machines), where the winning objective is specified by a regular set of target configurations and a qualitative probability constraint ‘>0’ or ‘=1’. The goal of one player is to maximize the probability of reaching the target set so that the constraint is satisfied, while the other player aims at the opposite. We show that the winner in such games can be determined in P for the ‘>0’ constraint, and in NP∩co-NP for the ‘=1’ constraint. Further, we prove that the winning regions for both players are regular, and we design algorithms which compute the associated finite-state automata. Finally, we show that winning strategies can be synthesized effectively.     

Tracking Context Changes through Meta-Learning

The article deals with the problem of learning incrementally (on-line) in domains where the target concepts are context-dependent, so that changes in context can produce more or less radical changes in the associated concepts. In particular, we concentrate on a class of learning tasks where the domain provides explicit clues as to the current context (e.g., attributes with characteristic values). A general two-level learning model is presented that effectively adjusts to changing contexts by trying to detect (via meta-learning) contextual clues and using this information to focus the learning process. Context learning and detection occur during regular on-line learning, without separate training phases for context recognition. Two operational systems based on this model are presented that differ in the underlying learning algorithm and in the way they use contextual information: METAL(B) combines meta-learning with a Bayesian classifier, while METAL(IB) is based on an instance-based learning algorithm. Experiments with synthetic domains as well as a number of real-world problems show that the algorithms are robust in a variety of dimensions, and that meta-learning can produce substantial increases in accuracy over simple object-level learning in situations with changing contexts.     

On the Input Problem for Massive Modularity

Jerry Fodor argues that the massive modularity thesis – the claim that (human) cognition is wholly served by domain specific, autonomous computational devices, i.e., modules – is a priori incoherent, self-defeating. The thesis suffers from what Fodor dubs the input problem: the function of a given module (proprietarily understood) in a wholly modular system presupposes non-modular processes. It will be argued that massive modularity suffers from no such a priori problem. Fodor, however, also offers what he describes as a really real input problem (i.e., an empirical one). It will be suggested that this problem is real enough, but it does not selectively strike down massive modularity – it is a problem for everyone.     

Emergence and Categorization of Coordinated Visual Behavior Through Embodied Interaction

This paper discusses the emergence of sensorimotor coordination for ESCHeR, a 4DOF redundant foveated rob ot-head, by interaction with its environment. A feedback-error-learning(FEL)-based distributed control provides the system with explorative abilities with reflexes constraining the learning space. A Kohonen network, trained at run-time, categorizes the sensorimotor patterns obtained over ESCHeR's interaction with its environment, enables the reinforcement of frequently executed actions, thus stabilizing the learning activity over time. We explain how the development of ESCHeR's visual abilities (namely gaze fixation and saccadic motion), from a context-free reflex-based control process to a context-dependent, pattern-based sensorimotor coordination can be related to the Piagetian stage theory.     

Barriers in the selection of offshore software development outsourcing vendors: An exploratory study using a systematic literature review

Context

Software development outsourcing is a contract-based relationship between client and vendor organisations in which a client contracts out all or part of its software development activities to a vendor, who provides agreed services for remuneration.

Objective

The objective is to identify various barriers that have a negative impact on software outsourcing clients in the selection process of offshore software development outsourcing vendors.

Method

We have performed a systematic literature review (SLR) process for the identification of barriers. We have performed all the SLR steps such as the protocol development, initial selection, final selection, quality assessment, data extraction and data synthesis.

Results

We have identified barriers such as ‘language and cultural barriers’, ‘country instability’, ‘lack of project management’, ‘lack of protection for intellectual property rights’ and ‘lack of technical capability’ that generally have a negative impact on outsourcing clients. We have identified only one common frequently cited barrier in three types of organisations (i.e. small, medium and large) which is ‘language and cultural barriers’. We did not identify any common frequently cited barrier in three continents (Asia, North America and Europe) and in two decades (1990-1999 and 2000-mid 2008). The results also reveal the similarities and differences in the barriers identified through different study strategies.

Conclusions

Vendors should address frequently cited barriers such as ‘language and cultural barriers’, ‘country instability’, ‘lack of project management’, ‘lack of protection for intellectual property rights’ and ‘lack of technical capability’ in order to compete in the offshore outsourcing business.     

Strategies for improving neural net generalisation

We address the problem of training multilayer perceptrons to instantiate a target function. In particular, we explore the accuracy of the trained network on a test set of previously unseen patterns — the generalisation ability of the trained network. We systematically evaluate alternative strategies designed to improve the generalisation performance. The basic idea is to generate a diverse set of networks, each of which is designed to be an implementation of the target function. We then have a set of trained, alternative versions — a version set. The goal is to achieve useful diversity within this set, and thus generate potential for improved generalisation performance of the set as a wholewhen compared to the performance of any individual version. We define this notion of useful diversity, we define a metric for it, we explore a number of ways of generating it, and we present the results of an empirical study of a number of strategies for exploiting it to achieve maximum generalisation performance. The strategies encompass statistical measures as well as a selectornet approach which proves to be particularly promising. The selector net is a form of metanet that operates in conjunction with a version set.     

Complex-Weighted One-Unit ‘Rigid-Bodies’ Learning Rule for Independent Component Analysis

Over the recent years, noticeable theoretical efforts have been devoted to the understanding of the role of networks' parameter spaces in neural learning. One of the contributions in this field concerns the study of weight-flows on Stiefel manifold, which is the natural parameter-space's algebraic-structure in some unsupervised (information-theoretic) learning task. An algorithm belonging to the class of learning equations generating Stiefel-flows is based on the rigid-body theory, introduced by the present Author in 1996. The aim of this Letter is to present an investigation on the capability of a complex-weighted neuron, trained by a rigid-bodies learning theory, with application to blind source separation of complex-valued independent signals for telecommunication systems.     

Coordinating Multiple Agents via Reinforcement Learning

In this paper, we attempt to use reinforcement learning techniques to solve agent coordination problems in task-oriented environments. The Fuzzy Subjective Task Structure model (FSTS) is presented to model the general agent coordination. We show that an agent coordination problem modeled in FSTS is a Decision-Theoretic Planning (DTP) problem, to which reinforcement learning can be applied. Two learning algorithms, coarse-grained and fine-grained, are proposed to address agents coordination behavior at two different levels. The coarse-grained algorithm operates at one level and tackle hard system constraints, and the fine-grained at another level and for soft constraints. We argue that it is important to explicitly model and explore coordination-specific (particularly system constraints) information, which underpins the two algorithms and attributes to the effectiveness of the algorithms. The algorithms are formally proved to converge and experimentally shown to be effective.     

Emergence and Categorization of Coordinated Visual Behavior Through Embodied Interaction

This paper discusses the emergence of sensorimotor coordination for ESCHeR, a 4DOF redundant foveated robot-head, by interaction with its environment. A feedback-error-learning (FEL)-based distributed control provides the system with explorative abilities with reflexes constraining the learning space. A Kohonen network, trained at run-time, categorizes the sensorimotor patterns obtained over ESCHeR's interaction with its environment, enables the reinforcement of frequently executed actions, thus stabilizing the learning activity over time. We explain how the development of ESCHeR's visual abilities (e.g., gaze fixation and saccadic motion), from a context-free reflex-based control process to a context-dependent, pattern-based sensorimotor coordination can be related to the Piagetian stage theory.     

Superpositional connectionism: A reply to Marinov

Marinov's critique I argue, is vitiated by its failure to recognize the distinctive role of superposition within the distributed connectionist paradigm. The use of so-called subsymbolic distributed encodings alone is not, I agree, enough to justify treating distributed connectionism as a distinctive approach. It has always been clear that microfeatural decomposition is both possible and actual within the confines of recognizably classical approaches. When such approaches also involve statistically-driven learning algorithms — as in the case of ID3 — the fundamental differences become even harder to spot. To see them, it is necessary to consider not just the nature of an acquired input-output function but the nature of the representational scheme underlying it. Differences between such schemes make themselves best felt outside the domain of immediate problem solving. It is in the more extended contexts of performance DURING learning and cognitive change as a result of SUBSEQUENT training on new tasks (or simultaneous training on several tasks) that the effects of superpositional storage techniques come to the fore. I conclude that subsymbols, distribution and statistically driven learning alone are indeed not of the essence. But connectionism is not just about subsymbols and distribution. It is about the generation of whole subsymbol SYSTEMS in which multiple distributed representations are created and superposed.Some of the material in sections 2 and 3 is drawn from Clark, A., ASSOCIATIVE ENGINES: CONNECTIONISM, CONCEPTS AND REPRESENTATIONAL CHANGE (Cambridge, MA MIT/Bradford Books, forthcoming 1993). Thanks to the publishers for permission to reproduce that material here. Thanks also to Chris Thornton for an illuminating discussion of ID3.     

Pressure-induced structural change from low-density to high-density amorphous ice by molecular-dynamics simulations

We have investigated the pressure dependence of the structure of amorphous ices by molecular-dynamics simulations to clarify the characteristic features of the structural change from the low-density amorphous (LDA) ice to the high-density amorphous (HDA) ice. We have found that, with increasing pressure at the temperature of 77 K, (1) the mean-square-displacement and its fluctuation as a function of time decrease, (2) the second peak of the oxygen–oxygen (OO) correlation function g(r) shifts to shorter distances and changes its shape drastically, e.g. it splits into two peaks, though the first peak remains almost same, (3) the average coordination number estimated from the first peak of g(r) increases from 4 to 5, (4) the broad peak of the OOO bond-angle distribution around 110° decreases and shifts towards 60°. From these results we conclude that the four-fold coordinated tetrahedral local structure in LDA ice changes to the five-fold coordinated structure in HDA ice, where the ‘fifth’ neighboring water molecule corresponds to the ‘interstitial’ water molecule.     

The use of ‘Clickers’ in the classroom: Teaching innovation or merely an amusing novelty?

‘Clickers’ are individual response devices in which students each have a remote control that allows them to quickly and anonymously respond to questions presented in-class. Clickers are now being used in many classrooms as an active learning component of courses. Educators considering the use of clickers in their own classrooms may wonder whether the clickers are a worthwhile, pedagogical tool or merely an amusing novelty. As Li (2008) pointed out, research has examined clicker effects on interaction within the classroom, but little research has examined whether clicker use can affect the understanding of concepts. This article will discuss ways in which clickers may help students organize and understand material presented in the classroom. The paper is intended to help guide educators in potentially effective uses of clickers as well as to guide future research.     

Collaboration and Teamwork: Immersion and Presence in an Online Learning Environment

In the world of OTIS, an online Internet School for occupational therapists, students from four European countries were encouraged to work collaboratively through problem-based learning by interacting with each other in a virtual semi-immersive environment. This paper describes, often in their own words, the experience of European occupational therapy students working together across national and cultural boundaries. Collaboration and teamwork were facilitated exclusively through an online environment, since the students never met each other physically during the OTIS pilot course. The aim of the paper is to explore the observations that (1) there was little interaction between students from different tutorial groups and (2) virtual teamwork developed in each of the cross-cultural tutorial groups. Synchronous data from the students was captured during tutorial sessions and peer-booked meetings and analyzed using the qualitative constructs of immersion, presence and reflection in learning. The findings indicate that immersion was experienced only to a certain extent. However, students found both presence and shared presence, within their tutorial groups, to help collaboration and teamwork. Other evidence suggests that communities of interest were established. Further study is proposed to support group work in an online learning environment. It is possible to conclude that collaborative systems can be designed, which encourage students to build trust and teamwork in a cross cultural online learning environment.This revised version was published online in March 2005 with corrections to the cover dateFunded by the European Union through the TENTelecom programme.     

Sparse Support Vector Machine with <Emphasis Type="Italic">L</Emphasis><Subscript><Emphasis Type="Italic">p</Emphasis></Subscript> Penalty for Feature Selection

We study the strategies in feature selection with sparse support vector machine (SVM). Recently, the socalled L _p -SVM (0 < p < 1) has attracted much attention because it can encourage better sparsity than the widely used L ₁-SVM. However, L _p -SVM is a non-convex and non-Lipschitz optimization problem. Solving this problem numerically is challenging. In this paper, we reformulate the L _p -SVM into an optimization model with linear objective function and smooth constraints (LOSC-SVM) so that it can be solved by numerical methods for smooth constrained optimization. Our numerical experiments on artificial datasets show that LOSC-SVM (0 < p < 1) can improve the classification performance in both feature selection and classification by choosing a suitable parameter p. We also apply it to some real-life datasets and experimental results show that it is superior to L ₁-SVM.     

The values of college students in business simulation game: A means-end chain approach

Business simulation games (BSGs) enable students to practice making decisions in a virtual environment, accumulate experience in application of strategies, and train themselves in modes of decision-making. This study examines the value sought by players of BSG. In this study, a means-end chain (MEC) model was adopted as the basis, and ladder method soft laddering was used to conduct in-depth interviews with students who had experience in using BSGs. The chain concept of “attribute-consequence-value” was used to understand students’ value cognition structures. Content analysis was used to analyze the attributes-consequences-values for BSGs players, then converted into a Hierarchical Value Map (HVM). The results showed that students consider teamwork and market diversity as the most important attributes, and the consequences of a cooperative approach and market diversity are emotional exchange and multi-thinking, with the ultimate value brought to users by exchanges between teams and constant thinking being interpersonal relationships and a sense of accomplishment.     

Using Multi-Layer Perceptrons to Predict Vehicle Pass-By Noise

All new vehicle designs have to pass a legislative, noise emission test – the pass-by noise test. In the highly competitive automotive industry, it is important to predict the test result early in the design process, rather than waiting until a prototype is built. Engineers can guess test results about aswell as the best, although inadequate, analytical models. They achieve this by using experience and their knowledge of acoustics and of the vehicles design. Neural networks should also be capable of pass-by noise prediction, learning from the results of previous tests. This paper describes a neural network approach to the problem. First, expert knowledge is used to select vehicle design and test parameters to present as inputs to a multi-layer perceptron. Since data is scarce, the problem is broken down into two stages, vehicle performance and pass-by noise. The two trained networks are evaluated and their performance discussed.     

The STEP Integration Information Architecture

Traditionally, when the term integration is used to refer to the interoperability of disparate, heterogeneous computer systems, it means the ability to exchange digital data between the systems. For the more sophisticated systems designers, integration may mean shared, distributed databases or a federated database system. Within the development of the STandard for the Exchange of Product model data (STEP — ISO 10303), integration refers to an information architecture composed of conceptual constructs that is independent of implementation considerations.The Integration Information Architecture of STEP is presented and explained. Instead of a flat representation of abstract (i.e., conceptual) data structures, integration within STEP takes place at four different levels:

Brave Mobots Use Representation: Emergence of Representation in Fight-or-Flight Learning

The paper uses ideas from Machine Learning, Artificial Intelligence and Genetic Algorithms to provide a model of the development of a fight-or-flight response in a simulated agent. The modelled development process involves (simulated) processes of evolution, learning and representation development. The main value of the model is that it provides an illustration of how simple learning processes may lead to the formation of structures which can be given a representational interpretation. It also shows how these may form the infrastructure for closely-coupled agent/environment interaction.     

Cognition and the Power of Continuous Dynamical Systems

Traditional approaches to modeling cognitive systems are computational, based on utilizing the standard tools and concepts of the theory of computation. More recently, a number of philosophers have argued that cognition is too subtle or complex for these tools to handle. These philosophers propose an alternative based on dynamical systems theory. Proponents of this view characterize dynamical systems as (i) utilizing continuous rather than discrete mathematics, and, as a result, (ii) being computationally more powerful than traditional computational automata. Indeed, the logical possibility of such super-powerful systems has been demonstrated in the form of analog artificial neural networks. In this paper I consider three arguments against the nomological possibility of these automata. While the first two arguments fail, the third succeeds. In particular, the presence of noise reduces the computational power of analog networks to that of traditional computational automata, and noise is a pervasive feature of information processing in biological systems. Consequently, as an empirical thesis, the proposed dynamical alternative is under-motivated: What is required is an account of how continuously valued systems could be realized in physical systems despite the ubiquity of noise.     

Program in C for studying characteristic properties of two-body interactions in the framework of spectral distribution theory

We present a program in C that employs spectral distribution theory for studies of characteristic properties of a many-particle quantum-mechanical system and the underlying few-body interaction. In particular, the program focuses on two-body nuclear interactions given in a JT$JT$-coupled harmonic oscillator basis and calculates correlation coefficients, a measure of similarity of any two interactions, as well as Hilbert–Schmidt norms specifying interaction strengths. An important feature of the program is its ability to identify the monopole part (centroid) of a 2-body interaction, as well as its ‘density-dependent’ one-body and two-body part, thereby providing key information on the evolution of shell gaps and binding energies for larger nuclear systems. As additional features, we provide statistical measures for ‘density-dependent’ interactions, as well as a mechanism to express an interaction in terms of two other interactions. This, in turn, allows one to identify, e.g., established features of the nuclear interaction (such as pairing correlations) within a general Hamiltonian. The program handles the radial degeneracy for ‘density-dependent’ one-body interactions and together with an efficient linked list data structure, facilitates studies of nuclear interactions in large model spaces that go beyond valence-shell applications.