Hybrid data-driven models of machine translation

This paper presents an extended, harmonised account of our previous work on combining subsentential alignments from phrase-based statistical machine translation (SMT) and example-based MT (EBMT) systems to create novel hybrid data-driven systems capable of outperforming the baseline SMT and EBMT systems from which they were derived. In previous work, we demonstrated that while an EBMT system is capable of outperforming a phrase-based SMT (PBSMT) system constructed from freely available resources, a hybrid ‘example-based’ SMT system incorporating marker chunks and SMT subsentential alignments is capable of outperforming both baseline translation models for French–English translation. In this paper, we show that similar gains are to be had from constructing a hybrid ‘statistical’ EBMT system. Unlike the previous research, here we use the Europarl training and test sets, which are fast becoming the standard data in the field. On these data sets, while all hybrid ‘statistical’ EBMT variants still fall short of the quality achieved by the baseline PBSMT system, we show that adding the marker chunks to create a hybrid ‘example-based’ SMT system outperforms the two baseline systems from which it is derived. Furthermore, we provide further evidence in favour of hybrid systems by adding an SMT target-language model to the EBMT system, and demonstrate that this too has a positive effect on translation quality. We also show that many of the subsentential alignments derived from the Europarl corpus are created by either the PBSMT or the EBMT system, but not by both. In sum, therefore, despite the obvious convergence of the two paradigms, the crucial differences between SMT and EBMT contribute positively to the overall translation quality. The central thesis of this paper is that any researcher who continues to develop an MT system using either of these approaches will benefit further from integrating the advantages of the other model; dogged adherence to one approach will lead to inferior systems being developed.     

Modeling and Control of Linear Two-time Scale Systems: Applied to Single-Link Flexible Manipulator

This paper deals with the problem of H _∞ control of linear two-time scale systems. The authors’ attention is focused on the robust regulation of the system based on a new modeling approach under the assumption of norm-boundedness of the fast dynamics. In the proposed approach, the fast dynamics are treated as a norm-bounded disturbance (dynamic uncertainty). In this view, the synthesis is performed only for the certain dynamics of the two-time scale system, whose order is less than that of the original system. It should be noted, however, that this scheme is significantly different from the conventional approaches of order reduction for linear two-time scale systems. Specifically, in the present work, explicitly or implicitly, all the dynamics of the system are taken into consideration. In other words, the portion that is treated as a perturbation is incorporated in the design by its maximum possible gain – in the L ₂ sense – over different values of the inputs. One of the advantagesof this approach is that – unlike in the conventional approaches of the order reduction – the reduced-order system still keeps some information of the ‘deleted’ subsystem. Also, we consider the robust stability analysis and stability bound improvement of perturbed parameter (ɛ) in the two-time scale systems by using linear fractional transformations and structured singular values (μ) approach. In this direction, by introducing the parametric uncertainty and dynamic uncertainty in the two-time scale systems, we represent the system as a standard μ-interconnection framework by using linear fractional transformations, and derive a set of new stability conditions for the system in the frequency domain. The exact solution of ɛ-bound is characterized. It is shown that, in spite of the coupling between the dynamic uncertainties and certain dynamics, the designed H _∞ controller stabilizes the overall closed-loop system, in the presence of norm-bounded disturbances. To show the effectiveness of the approach, the modeling of the single-link flexible manipulator and control of the Tip-position of the manipulator by utilizing the mentioned method are presented in the case study.     

The Context of Information Use in a Hospital as Simultaneous Similarity–Difference Relations

In the light of the developing discourse on the relative merits of ‘hard’ and ‘soft’ approaches to information systems development, we present a case study application of a methodology which attempts to dissolve such dualities. Personal Construct Psychology (PCP) offers, as a unity, the construing person who is both biology and culture. PCP argues that both the world and the person’s construct system are phenomenologically real and that the viability of any particular construct system depends only on its usefulness to the construing person. In this study, we used PCP to explore the organisational context of information use and distribution in a large hospital. We used repertory grids, a PCP technique, to elicit from 16 members of staff their personal construals of information from different sources in the hospital. The results highlight the relationship between meaningful information and meaningfully active relationships, a theme which we discuss in terms of the development of the hospital information system and in terms of the value of PCP in dissolving hard–soft dichotomies.     

Interaction effects in simultaneous motor control and movement perception tasks

Recent findings in neuroscience suggest an overlap between those brain regions involved in the control and execution of movement and those activated during the perception of another’s movement. This so called ‘mirror neuron’ system is thought to underlie our ability to automatically infer the goals and intentions of others by observing their actions. Kilner et al. (Curr Biol 13(6):522–525, 2003) provide evidence for a human ‘mirror neuron’ system by showing that the execution of simple arm movements is affected by the simultaneous perception of another’s movement. Specifically, observation of ‘incongruent’ movements made by another human, but not by a robotic arm, leads to greater variability in the movement trajectory than observation of movements in the same direction. In this study we ask which aspects of the observed motion are crucial to this interference effect by comparing the efficacy of real human movement to that of sparse ‘point-light displays’. Eight participants performed whole arm movements in both horizontal and vertical directions while observing either the experimenter or a virtual ‘point-light’ figure making arm movements in the same or in a different direction. Our results, however, failed to show an effect of ‘congruency’ of the observed movement on movement variability, regardless of whether a human actor or point-light figure was observed. The findings are discussed, and future directions for studies of perception-action coupling are considered.     

Non-Uniform Reductions

We study properties of non-uniform reductions and related completeness notions. We strengthen several results of Hitchcock and Pavan (ICALP (1), Lecture Notes in Computer Science, vol. 4051, pp. 465–476, Springer, 2006) and give a trade-off between the amount of advice needed for a reduction and its honesty on NEXP. We construct an oracle relative to which this trade-off is optimal. We show, in a more systematic study of non-uniform reductions, among other things that non-uniformity can be removed at the cost of more queries. In line with Post’s program for complexity theory (Buhrman and Torenvliet in Bulletin of the EATCS 85, pp. 41–51, 2005) we connect such ‘uniformization’ properties to the separation of complexity classes.     

‘Working Well, Together’: Arts-Based Research and the Cultural Future of Small Cities

This article offers a research update on a 3-year programme initiated by the Kamloops Art Gallery and the University College of the Cariboo in Kamloops, British Columbia. The programme is supported by a ‘Community–University Research Alliance’ grant from the Social Sciences and Humanities Research Council of Canada, and the collaboration focuses on the cultural future of small cities – on how cultural and arts organisations work together (or fail to work together) in a small city setting. If not by definition, then certainly by default, ‘culture’ is associated with big city life: big cities are equated commonly with ‘big culture’; small cities with something less. The Cultural Future of Small Cities research group seeks to provide a more nuanced view of what constitutes culture in a small Canadian city. In particular, the researchers are exploring notions of social capital and community asset building: in this context, ‘visual and verbal representation’, ‘home’, ‘community’ and the need to define a local ‘sense of place’ have emerged as important themes. As the Small Cities programme begins its second year, a unique but key aspect has become the artist-as-researcher. Correspondence and offprint requests to: L. Dubinsky, Kamloops Art Gallery, 101–465 Victoria Street, Kamloops, BC V2C 2A9 Canada. Tel.: 250-828-3543; Email: ldubinsky@museums.ca     

Efficient distribution-free population learning of simple concepts

We consider a variant of the ‘population learning model’ proposed by Kearns and Seung [8], in which the learner is required to be ‘distribution-free’ as well as computationally efficient. A population learner receives as input hypotheses from a large population of agents and produces as output its final hypothesis. Each agent is assumed to independently obtain labeled sample for the target concept and output a hypothesis. A polynomial time population learner is said to PAC-learn a concept class, if its hypothesis is probably approximately correct whenever the population size exceeds a certain bound which is polynomial, even if the sample size for each agent is fixed at some constant. We exhibit some general population learning strategies, and some simple concept classes that can be learned by them. These strategies include the ‘supremum hypothesis finder’, the ‘minimum superset finder’ (a special case of the ‘supremum hypothesis finder’), and various voting schemes. When coupled with appropriate agent algorithms, these strategies can learn a variety of simple concept classes, such as the ‘high–low game’, conjunctions, axis-parallel rectangles and others. We give upper bounds on the required population size for each of these cases, and show that these systems can be used to obtain a speed up from the ordinary PAC-learning model [11], with appropriate choices of sample and population sizes. With the population learner restricted to be a voting scheme, what we have is effectively a model of ‘population prediction’, in which the learner is to predict the value of the target concept at an arbitrarily drawn point, as a threshold function of the predictions made by its agents on the same point. We show that the population learning model is strictly more powerful than the population prediction model. Finally, we consider a variant of this model with classification noise, and exhibit a population learner for the class of conjunctions in this model. This revised version was published online in June 2006 with corrections to the Cover Date.     

An Automatic Correction of Ma’s Thinning Algorithm Based on <Emphasis Type="Italic">P</Emphasis>-simple Points

The notion of P-simple points was introduced by Bertrand to conceive parallel thinning algorithms. In ‘A 3D fully parallel thinning algorithm for generating medial faces’ (Pattern Recogn. Lett. 16:83–87, 1995), Ma proposed an algorithm for which there are objects whose topology is not preserved. In this paper, we propose a new application of P-simple points: to automatically correct Ma’s algorithm.     

Decision Making and Management of Dynamic Risk

Managing dynamic environments often requires decision making under uncertainty and risk. Two types of uncertainty are involved: uncertainty about the state and the evolution of the situation, and ‘openness’ of the possible actions to face possible consequences. In an experimental study on risk management in dynamic situations, two contrasted ‘ecological’ scenarios – transposed from effective situations of emergency management – were compared in order to identify the impact of their ‘openness’ in the subjects’ strategies for decision making. The ‘Lost Child’ scenario presented qualitative and irreversible consequences (child’s death) and high uncertainty; it exerted high demands both in risk assessment (risk representation) and action elaboration and choice. A less open situation (‘Hydrocarbon Fire’) required a main choice between two contrasted actions, with quantitative computable consequences. The strategies of ‘experimental subjects’ (university students) and ‘operative subjects’ (professional fire-fighter officers) were compared in order to evaluate the ecological validity of experimental research in this field, from the point of view of the subjects themselves. The two scenarios appeared to be independent, so that quite different models of decision making have to be hypothesised, differing by the importance of assessing risk and defining possible actions on the one hand, and by the process of choice on the other. ‘Experimental’ subjects dramatically differed from ‘operative’ subjects when confronted with the same scenario, particularly for the less technical but more demanding scenario. It is hypothesised that three components might account for the effect of the situations and for the differences between and within groups of subjects: importance of situation assessment, spatial abilities, and global orientation of activity in managing dynamic risk.     

Pest Control Expert System for Tomato (PCEST)

This paper presents a real-life pest control expert system for tomato. The system involves two main subtasks, namely: ‘diagnose’ and ‘treat’. The ‘diagnose’ subtask finds out the causes of the growers' complaints, while the ‘treat’ subtask finds out a treatment plan for these causes. CommonKADS methodology has been used to develop the system. Dependency network is used as one of our knowledge representation schemes in both subtasks. An expert system evaluation methodology has been suggested and applied to the developed system. Received May 1998 / Revised January 1999 / Accepted in revised form May 1999     

The Cooperative Work of Gaming: Orchestrating a Mobile SMS Game

This paper focuses on orchestration work in the first iteration of a mobile game called Day Of The Figurines, which explores the potential to exploit text messaging as a means of creating an engaging gaming experience. By focusing on orchestration we are especially concerned with the ‘cooperative work that makes the game work’. While the assemblage or family of orchestration practices uncovered by our ethnographic study are specific to the game – including the ways in which behind the scenes staff make sense of messages, craft appropriate responses, and manage and track the production of gameplay narratives as the game unfolds – orchestration work is of general significance to our understanding of new gaming experiences. The focus on orchestration work reveals that behind the scenes staff are co-producers of the game and that the playing of games is, therefore, inseparably intertwined with their orchestration. Furthermore, orchestration work is ‘ordinary’ work that relies upon the taken for granted skills and competences of behind the scenes staff; ‘operators’ and ‘authors’ in this case. While we remain focused on the specifics of this game, explication of the ordinary work of orchestration highlights challenges and opportunities for the continued development of gaming experiences more generally. Indeed, understanding the specificities of orchestration work might be said to be a key ingredient of future development.     

How Can I Help You? Call Centres,Classification Work and Coordination

As a comparatively novel but increasingly pervasive organizational arrangement, call centres have been a focus for much recent research. This paper identifies lessons for organizational and technological design through an examination of call centres and ‘classification work’ – explicating what Star [1992, Systems/Practice vol. 5, pp. 395–410] terms the ‘open black box’. Classification is a central means by which organizations standardize procedure, assess productivity, develop services and re-organize their business. Nevertheless, as Bowker and Star [1999, Sorting Things Out: Classification and Its Consequences. Cambridge MA: MIT Press] have pointed out, we know relatively little about the work that goes into making classification schema what they are. We will suggest that a focus on classification ‘work’ in this context is a useful exemplar of the need for some kind of ‘meta-analysis’ in ethnographic work also. If standardization is a major ambition for organizations under late capitalism, then comparison might be seen as a related but as-yet unrealized one for ethnographers. In this paper, we attempt an initial cut at a comparative approach, focusing on classification because it seemed to be the primary issue that emerged when we compared studies. Moreover, if technology is the principal means through which procedure and practice is implemented and if, as we believe, classifications are becoming ever more explicitly embedded within it (for instance with the development of so-called ‘semantic web’ and associated approaches to ontology-based design), then there is clearly a case for identifying some themes which might underpin classification work in a given domain.     

Antisedentary Beigeless Computing

Conclusion Four decades of sporadic invention and experimentation of and with non-traditional human-computer interface schemes have congealed (somewhat abruptly though not without a few clear-sighted antecedents) into a new field of information system design, here calledAntisedentary Beigeless Computing, that consciously rejects the traditional conception of isolated tete-a-tete between the human and the box-CRT-keyboardmouse. ABC systems instead favour the complementary directions away from this notion of an immobile info-shrine: more personal, intimate, and portable information access; and more diffuse, environmentally-integrated information access. Consideration of ABC projects to date seems to suggest that no single instance can alone express the full generality required of a ‘working’ information system, so that (on the one hand) system design must acknowledge that a complex set of trade-offs involving capabilities, universality, specificity, personalization, and generality is inescapable; while (on the other hand) an ideal, eventual ‘information environment’ will inevitably comprise the careful interweaving of some number of individual ABC systems. Taxonomies and classification schema can rarely hope to be found complete or flawless before the collection of items that they purport to describe have themselves reached the evolutionary stasis of ‘adulthood’ — that is, there is typically some threshold of development or growth beyond which few enough surprises lurk that an encompassing taxonomy can be constructed and observed to reliably encompass, in the longer term. The domain of ABC thought is still quite nascent, and so we would be foolish to assume that all its extremities of form and connotation are now visible, but to the extent that we can already see the outlines of a ‘field’ it is reasonable to make a first run at an analytic taxonomy. The ‘independent character axes’ approach presented here seems broad and loose enough to accommodate any number of additions to the basic stable of ABC systems. It is, further, a taxonomy amenable to significant revision as may be found necessary: axes can be added, deleted, reconstrued, etc. as time and consideration clarify our understanding of ABC. However, it should also be anticipated that the field will eventually coalesce around a much smaller number of better-defined ‘axes’ and thus permit taxonomic reversion to the more hierarchical (and finally more satisfying) ‘Linnean’ scheme we'd originally imagined establishing.     

The significant other: the value of jewellery in the conception,design and experience of body focused digital devices

In this paper, we demonstrate how craft practice in contemporary jewellery opens up conceptions of ‘digital jewellery’ to possibilities beyond merely embedding pre-existing behaviours of digital systems in objects, which follow shallow interpretations of jewellery. We argue that a design approach that understands jewellery only in terms of location on the body is likely to lead to a world of ‘gadgets’, rather than anything that deserves the moniker ‘jewellery’. In contrast, by adopting a craft approach, we demonstrate that the space of digital jewellery can include objects where the digital functionality is integrated as one facet of an object that can be personally meaningful for the holder or wearer.     

DSM as a knowledge capture tool in CODE environment

A design structure matrix (DSM) provides a simple, compact, and visual representation of a complex system/ process. This paper shows how DSM, a system engineering tool, is applied as a knowledge capture (acquisition) tool in a generic NPD process. The acquired knowledge (identified in the DSM) is provided in the form of Questionnaires, which are organized into five performance indicators of the organization namely ‘Marketing’, ‘Technical’, ‘Financial’, ‘Resource Management’, and ‘Project Management’. Industrial application is carried out for knowledge validation. It is found form the application that the acquired knowledge helps NPD teams, managers and stakeholders to benchmark their NPD endeavor and select areas to focus their improvement efforts (up to 80% valid).     

Coverage-guided test generation for continuous and hybrid systems

In this paper, we describe a formal framework for conformance testing of continuous and hybrid systems, using the international standard ‘Formal Methods in Conformance Testing’ FMCT. We propose a novel test coverage measure for these systems, which is defined using the star discrepancy notion. This coverage measure is used to quantify the validation ‘completeness’. It is also used to guide input stimulus generation by identifying the portions of the system behaviors that are not adequately examined. We then propose a test generation method, which is based on a robotic motion planning algorithm and is guided by the coverage measure. This method was implemented in a prototype tool that can handle high dimensional systems (up to 100 dimensions).     

Stability of solutions of delay functional integro-differential equations and their discretizations

In this paper we study asymptotic stability and contractivity properties of solutions of a class of delay functional integro-differential equations. These results form the basis for obtaining insight into the analogous properties of numerical solutions generated by continuous Runge-Kutta or collocation methods, where these methods are applied to a suitable reformulation of the given initial-value problem.     

Using Dempster–Shafer to incorporate knowledge into satellite image classification

Remote sensing imaging techniques make use of data derived from high resolution satellite sensors. Image classification identifies and organises pixels of similar spatial distribution or similar statistical characteristics into the same spectral class (theme). Contextual data can be incorporated, or ‘fused’, with spectral data to improve the accuracy of classification algorithms. In this paper we use Dempster–Shafer’s theory of evidence to achieve this data fusion. Incorporating a Knowledge Base of evidence within the classification process represents a new direction for the development of reliable systems for image classification and the interpretation of remotely sensed data.     

MABA-MABA or Abracadabra? Progress on Human–Automation Co-ordination

In this paper we argue that substitution-based function allocation methods (such as MABA-MABA, or Men-Are-Better-At/Machines-Are-Better-At lists) cannot provide progress on human–automation co-ordination. Quantitative ‘who does what’ allocation does not work because the real effects of automation are qualitative: it transforms human practice and forces people to adapt their skills and routines. Rather than re-inventing or refining substitution-based methods, we propose that the more pressing question on human–automation co-ordination is ‘How do we make them get along together?’ Correspondence and offprint requests to: S. W. A. Dekker, Department of Mechanical Engineering, IKP, Link?ping Institute of Technology, SE - 581 83 Link?ping, Sweden. Tel.: +46 13 281646; fax +4613282579; email: sidde@ikp.liu.se     

Quotient convergence and multi-splitting methods for solving singular linear equations

Abstract   In this paper, we use the group inverse to characterize the quotient convergence of an iterative method for solving consistent singular linear systems, when the matrix index equals one. Next, we show that for stationary splitting iterative methods, the convergence and the quotient convergence are equivalent, which was first proved in [7]. Lastly, we propose a (multi-)splitting iterative method A=F–G, where the splitting matrix F may be singular, endowed with group inverse, by using F ^# as a solution tool for any iteration. In this direction, sufficient conditions for the quotient convergence of these methods are given. Then, by using the equivalence between convergence and quotient convergence, the classical convergence of these methods is proved. These latter results generalize Cao’s result, which was given for nonsingular splitting matrices F. Keywords: Group inverse, singular linear equations, iterative method, P-regular splitting, Hermitian positive definite matrix, multi-splitting, quotient convergence AMS Classification: 15A09, 65F35