Permanent URI for this collection
Browse
Recent Submissions
Publication Dissimilarity-based representations for one-class classification on time series(Elsevier, 2019-11-25) Mauceri, Stefano; Sweeney, James; McDermott, James; ICON plc.In several real-world classification problems it can be impractical to collect samples from classes other than the one of interest, hence the need for classifiers trained on a single class. There is a rich literature concerning binary and multi-class time series classification but less concerning one-class learning. In this study, we investigate the little-explored one-class time series classification problem. We represent time series as vectors of dissimilarities from a set of time series referred to as prototypes. Based on this approach, we evaluate a Cartesian product of 12 dissimilarity measures, and 8 prototype methods (strategies to select prototypes). Finally, a one-class nearest neighbor classifier is used on the dissimilarity-based representations (DBR). Experimental results show that DBR are competitive overall when compared with a strong baseline on the data-sets of the UCR/UEA archive. Additionally, DBR enable dimensionality reduction, and visual exploration of data-sets.Publication Interpreting global perturbation robustness of image models using axiomatic spectral importance decomposition(Transactions on Machine Learning Research, 2024-07) Luo, Róisín (Jiaolin Luo); McDermott, James; O'Riordan, Colm; Hara, Satoshi; Science Foundation IrelandPerturbation robustness evaluates the vulnerabilities of models, arising from a variety of perturbations, such as data corruptions and adversarial attacks. Understanding the mechanisms of perturbation robustness is critical for global interpretability. We present a model-agnostic, global mechanistic interpretability method to interpret the perturbation robustness of image models. This research is motivated by two key aspects. First, previous global interpretability works, in tandem with robustness benchmarks, e.g., mean corruption error (mCE), are not designed to directly interpret the mechanisms of perturbation robustness within image models. Second, we notice that the spectral signal-to-noise ratios (SNR) of perturbed natural images exponentially decay over the frequency. This power-law-like decay implies that low-frequency signals are generally more robust than high-frequency signals—yet high classification accuracy cannot be achieved by low-frequency signals alone. By applying Shapley value theory, our method axiomatically quantifies the predictive powers of robust features and non-robust features within an information theory framework. Our method, dubbed as I-ASIDE (Image Axiomatic Spectral Importance Decomposition Explanation), provides a unique insight into model robustness mechanisms. We conduct extensive experiments over a variety of vision models pre-trained on ImageNet, including both convolutional neural networks (e.g., AlexNet, VGG, GoogLeNet/Inception-v1, Inception-v3, ResNet, SqueezeNet, RegNet, MnasNet, MobileNet, EfficientNet, etc.) and vision transformers (e.g., ViT, Swin Transformer, and MaxViT), to show that I-ASIDE can not only measure the perturbation robustness but also provide interpretations of its mechanisms.Publication Measuring node decentralisation in blockchain peer to peer networks(Elsevier, 2023-02-08) Howell, Andrew; Saber, Takfarinas; Bendechach, Malika; Science Foundation IrelandNew blockchain platforms are launching at a high cadence, each fighting for attention, adoption, and infrastructure resources. Several studies have measured the peer-to-peer (P2P) network decentralisation of Bitcoin and Ethereum (i.e., two of the largest used platforms). However, with the increasing demand for blockchain infrastructure, it is important to study node decentralisation across multiple blockchain networks, especially those containing a small number of nodes. In this paper, we propose NodeMaps, a data processing framework to capture, analyse, and visualise data from several popular P2P blockchain platforms, such as Cosmos, Stellar, Bitcoin, and Lightning Network. We compare and contrast the geographic distribution, the hosting provider diversity, and the software client variance in each of these platforms. Through our comparative analysis of node data, we found that Bitcoin and its Lightning Network Layer 2 protocol are widely decentralised P2P blockchain platforms, with the largest geographical reach and a high proportion of nodes operating on The Onion Router (TOR) privacy-focused network. Cosmos and Stellar blockchains have reduced node participation, with nodes predominantly operating in large cloud providers or well-known data centres.Publication Species mixing proportion and aridity influence in the height–diameter relationship for different species mixtures in Mediterranean forests(MDPI, 2022-01-14) Rodríguez de Prado, Diego; Riofrío, Jose; Aldea, Jorge; McDermott, James; Bravo, Felipe; Herrero de Aza, Celia; European Regional Development Fund; Horizon 2020Estimating tree height is essential for modelling and managing both pure and mixed forest stands. Although height–diameter (H–D) relationships have been traditionally fitted for pure stands, attention must be paid when analyzing this relationship behavior in stands composed of more than one species. The present context of global change makes also necessary to analyze how this relationship is influenced by climate conditions. This study tends to cope these gaps, by fitting new H–D models for 13 different Mediterranean species in mixed forest stands under different mixing proportions along an aridity gradient in Spain. Using Spanish National Forest Inventory data, a total of 14 height–diameter equations were initially fitted in order to select the best base models for each pair species-mixture. Then, the best models were expanded including species proportion by area (mi) and the De Martonne Aridity Index (M). A general trend was found for coniferous species, with taller trees for the same diameter size in pure than in mixed stands, being this trend inverse for broadleaved species. Regarding aridity influence on H–D relationships, humid conditions seem to beneficiate tree height for almost all the analyzed species and species mixtures. These results may have a relevant importance for Mediterranean coppice stands, suggesting that introducing conifers in broadleaves forests could enhance height for coppice species. However, this practice only should be carried out in places with a low probability of drought. Models presented in our study can be used to predict height both in different pure and mixed forests at different spatio-temporal scales to take better sustainable management decisions under future climate change scenarios.Publication A system dynamics approach to increasing ocean literacy(Frontiers Media, 2019-06-28) Green, Caroline; Ashley, Matthew; Molloy, Owen; Brennan, Caroline; Horizon 2020Ocean Literacy (OL) has multiple aspects or dimensions: from knowledge about how the oceans work and our impact on them, to attitudes toward topics such as sustainable fisheries, and our behaviour as consumers, tourists, policy makers, fishermen, etc. The myriad ways in which individuals, society and the oceans interact result in complex dynamic systems, composed of multiple interlinked chains of cause and effect. To influence our understanding of these systems, and thereby increase our OL, means to increase our knowledge of our own and others¿ place and role in the web of interactions. Systems Thinking has a potentially important role to play in helping us to understand, explain and manage problems in the human-ocean relationship. Leaders in the OL field have recommended taking a systems approach in order to deal with the complexity of the human-ocean relationship. They contend that the inclusion of modelling and simulation will improve the effectiveness of educational initiatives. In this paper we describe a pilot study centred on a browser-based Simulation-Based Learning Environment (SBLE) designed for a general audience that uses System Dynamics simulation to introduce and reinforce systems-based OL learning. It uses a storytelling approach, by explaining the dynamics of coastal tourism through a System Dynamics model revealed in stages, supported by fact panels, pictures, simulation-based tasks, causal loop diagrams and quiz questions. Participants in the pilot study were mainly postgraduate students. A facilitator was available to participants at all times, as needed. The model is based on a freely available normalised coastal tourism model by Hartmut Bossel, converted to XMILE format. Through the identification and use of systems archetypes and general systems features such as feedback loops, we also tested for the acquisition of transferable skills and the ability to identify, apply or create sustainable solutions. Levels of OL were measured before and after interaction with the tool using pre- and post-survey questionnaires and interviews. Results showed moderate to very large positive effects on all the OL dimensions, which are also shown to be associated with predictors of behaviour change. These results provide motivation for further research.Publication A system dynamics approach to sustainability education(Wiley, 2021-01-14) Green, Caroline; Molloy, Owen; Brennan, CarolineThis article describes a randomised controlled trial (RCT) study design to evaluate an innovative online tool to support education for sustainable development (ESD). The learning tool incorporates Systems Thinking and simulation into learning about two sustainability themes, deer herd management and fisheries. The study employs a factorial design to investigate whether Systems Thinking, simulation or both improve learning outcomes, when compared to a control group.Publication An empirical study of the impact of systems thinking and simulation on sustainability education(MDPI, 2021-12-30) Green, Caroline; Molloy, Owen; Duggan, Jim; Brennan, Caroline; Horizon 2020; Higher Education Authority; Department of Further and Higher Education; National University of Ireland, GalwayEducation for sustainable development (ESD) is considered vital to the success of the United Nations¿ sustainable development goals. Systems thinking has been identified as a core competency that must be included in ESD. However, systems thinking-orientated ESD learning tools, established methods of the assessment of sustainability skills, and formal trials to demonstrate the effectiveness of such learning tools are all lacking. This research presents a randomised controlled trial (n = 106) to investigate whether an innovative online sustainability learning tool that incorporates two factors, systems thinking and system dynamics simulation, increases the understanding of a specific sustainability problem. A further aim was to investigate whether these factors also support the transfer of knowledge to a second problem with a similar systemic structure. The effects of the two factors were tested separately and in combination using a two-by-two factorial study design. ANOVA and related inferential statistical techniques were used to analyse the effect of the factors on sustainability understanding. Cohen¿s d effect sizes were also calculated. Simulation alone was found to increase ESD learning outcomes significantly, and also to support the transfer of skills, although less significantly. Qualitative feedback was also gathered from participants, most of whom reported finding systems thinking and simulation very helpful.Publication Multi-objective multi-agent decision making: a utility-based analysis and survey(Springer Verlag, 2019-12-09) Rădulescu, Roxana; Mannion, Patrick; Roijers, Diederik M.; Nowé, AnnThe majority of multi-agent system implementations aim to optimise agents policies with respect to a single objective, despite the fact that many real-world problem domains are inherently multi-objective in nature. Multi-objective multi-agent systems (MOMAS) explicitly consider the possible trade-offs between conflicting objective functions. We argue that, in MOMAS, such compromises should be analysed on the basis of the utility that these compromises have for the users of a system. As is standard in multi-objective optimisation, we model the user utility using utility functions that map value or return vectors to scalar values. This approach naturally leads to two different optimisation criteria: expected scalarised returns (ESR) and scalarised expected returns (SER). We develop a new taxonomy which classifies multi-objective multi-agent decision making settings, on the basis of the reward structures, and which and how utility functions are applied. This allows us to offer a structured view of the field, to clearly delineate the current state-of-the-art in multi-objective multi-agent decision making approaches and to identify promising directions for future research. Starting from the execution phase, in which the selected policies are applied and the utility for the users is attained, we analyse which solution concepts apply to the different settings in our taxonomy. Furthermore, we define and discuss these solution concepts under both ESR and SER optimisation criteria. We conclude with a summary of our main findings and a discussion of many promising future research directions in multi-objective multi-agent systems.Publication When and why metaheuristics researchers can ignore “No Free Lunch” theorems(Springer, 2019-03-25) McDermott, JamesThe No Free Lunch (NFL) theorem for search and optimisation states that averaged across all possible objective functions on a fixed search space, all search algorithms perform equally well. Several refined versions of the theorem find a similar outcome when averaging across smaller sets of functions. This paper argues that NFL results continue to be misunderstood by many researchers, and addresses this issue in several ways. Existing arguments against real-world implications of NFL results are collected and re-stated for accessibility and new ones are added. Specific misunderstandings extant in the literature are identified, with speculation as to how they may have arisen. This paper presents an argument against a common paraphrase of NFL findings—that algorithms must be specialised to problem domains to do well—after problematising the usually undefined term “domain”. It provides novel concrete counter-examples illustrating cases where NFL theorems do not apply. In conclusion, it offers a novel view of the real meaning of NFL, incorporating the anthropic principle and justifying the position that in many common situations researchers can ignore NFL.Publication Input and output data analysis for system dynamics modelling using the tidyverse libraries of R(Wiley, 2018-09-07) Duggan, JimFrom a number of perspectives, system dynamics is a data‐intensive activity. First, each modelling challenge addresses behaviour over time, where historical time series data inform the model‐building process, and techniques such as calibration and optimisation (Rahmandad et al., 2015) are deployed to estimate parameters and enhance user confidence in model outputs. Second, the simulation of higher‐order models (Forrester, 1987) typically yields many time‐based observations across a significant number of variables. These results must be interpreted and analysed as part of the model‐building and policy analysis process. Third, simulation methods such as sensitivity analysis (Hekimoğlu and Barlas, 2016; Walrave, 2016; Jadun et al., 2017) generate large datasets that need to be processed for further analysis—for example, techniques such as statistical screening (Ford and Flynn, 2005; Taylor et al., 2010; Yasaman and Ford, 2016). Therefore, in the context of these data‐intensive modelling processes, there are opportunities for system dynamics modellers to leverage complementary data exploration technologies such as R (Duggan, 2016b). The R programming language provides a flexible framework for supporting system dynamics modelling. In particular, R now contains a suite of libraries, collectively known as the tidyverse,1 that are specifically designed to process rectangular data, which is highly structured data with rows as individual observations, and columns containing variables. In a system dynamics context, a row represents the simulation output at a unique point in time, and columns contain model variables (i.e., stocks, flows and auxiliaries). Given this perspective, the output from a system dynamics model with n time steps and m variables can be viewed as a single rectangular data set with dimensions (n × m). Many of the tidyverse libraries provide quick and efficient ways to process rectangular data, including importing data from external sources (e.g., comma‐separated files), summarising and aggregating observations (frequency counts, summary functions) and visualising large datasets. Before describing these libraries, an overview of R is provided.Publication Follow flee: A contingent mobility strategy for the spatial prisoner's dilemma(Springer Verlag, 2016-08-10) Gibbons, Maud D.; O'Riordan, Colm; Griffith, Josephine; Hardiman Research Scholarship, NUI GalwayThis paper presents results from a series of experimental simulations comparing the performances of mobile strategies of agents participating in the Spatial Prisoner's Dilemma game. The contingent movement strategies Walk Away and Follow Flee are evaluated and compared in terms of (1) their ability to promote the evolution of cooperation, and (2) their susceptibility to changes in the environmental and evolutionary settings. Results show that the Follow Flee strategy outperforms the Walk Away strategy across a broad range of environment parameter values, and exhibits the ability to invade the rival strategy. We propose that the Follow Flee movement strategy is successful due to its ability to pro-actively generate and maintain mutually cooperative relationships.Publication Simulation of an optional strategy in the prisoner’s dilemma in spatial and non-spatial environments(Springer Verlag, 2016-08-10) Cardinot, Marcos; Gibbons, Maud; O'Riordan, Colm; Griffith, Josephine; CNPq–Brazil; Hardiman Scholarship, NUI GalwayThis paper presents research comparing the effects of different environments on the outcome of an extended Prisoner's Dilemma, in which agents have the option to abstain from playing the game. We consider three different pure strategies: cooperation, defection and abstinence. We adopt an evolutionary game theoretic approach and consider two different environments: the first which imposes no spatial constraints and the second in which agents are placed on a lattice grid. We analyse the performance of the three strategies as we vary the loner's payoff in both structured and unstructured environments. Furthermore we also present the results of simulations which identify scenarios in which cooperative clusters of agents emerge and persist in both environments.Publication A further analysis of the role of heterogeneity in coevolutionary spatial games(Elsevier, 2017-11-04) Cardinot, Marcos; Griffith, Josephine; O'Riordan, Colm; National Council for Scientific and Technological Development (CNPq-Brazil)Heterogeneity has been studied as one of the most common explanations of the puzzle of cooperation in social dilemmas. A large number of papers have been published discussing the effects of increasing heterogeneity in structured populations of agents, where it has been established that heterogeneity may favour cooperative behaviour if it supports agents to locally coordinate their strategies. In this paper, assuming an existing model of a heterogeneous weighted network, we aim to further this analysis by exploring the relationship (if any) between heterogeneity and cooperation. We adopt a weighted network which is fully populated by agents playing both the Prisoner’s Dilemma or the Optional Prisoner’s Dilemma games with coevolutionary rules, i.e., not only the strategies but also the link weights evolve over time. Surprisingly, results show that the heterogeneity of link weights (states) on their own does not always promote cooperation; rather cooperation is actually favoured by the increase in the number of overlapping states and not by the heterogeneity itself. We believe that these results can guide further research towards a more accurate analysis of the role of heterogeneity in social dilemmas.Publication System dynamics modelling to support policy analysis for sustainable health care(2014) Lyons, Gerard J.; Duggan, JimSystem dynamics (SD) is an established simulation methodology used to explore the behaviour of social systems over time. The field has addressed challenging sustainability problems in fisheries, urban planning and environmental resource management. It has also been successfully applied to health care, in chronic disease modelling and workforce planning. This paper presents SD models of health-care sustainability, and illustrates two complementary applications of SD: (i) continuous simulation of health-care infrastructure adequacy; and (ii) conceptual modelling of the wider public policy context for health-care sustainability. The infrastructure model provides a simulator for evaluating impacts of population growth and ageing, as well as assessing the likely effects of policy interventions on system sustainability. This model is validated using empirical data from Ireland s public health service, and its practical application for sustainability analysis is illustrated. Our conceptual endogenous SD model explores a wider system boundary and public policy interdependencies that impact sustainability outcomes.Publication Improving spectral library search by redefining similarity measures(Journal Of Chemical Information And Modeling, 2015-04-22) Garg, Ankita; Enright, Catherine G.; Madden, Michael G.; |~|EU|~|Similarity plays a central role in spectral library search. The goal of spectral library search is to identify those spectra in a reference library of known materials that most closely match an unknown query spectrum, on the assumption that this will allow us to identify the main constituent(s) of the query spectrum. The similarity measures used for this task in software and the academic literature are almost exclusively metrics, meaning that the measures obey the three axioms of metrics: (1) minimality; (2) symmetry; (3) triangle inequality. Consequently, they implicitly assume that the query spectrum is drawn from the same distribution as that of the reference library.In this paper, we demonstrate that this assumption is not necessary in practical spectral library search and that in fact it is often violated in practice. Although the reference library may be constructed carefully, it is generally impossible to guarantee that all future query spectra will be drawn from the same distribution as the reference library. Before evaluating different similarity measures, we need to understand how they define the relationship between spectra.In spectral library search, we often aim to find the constituent(s) of a mixture. We propose that rather than asking which reference library spectra are similar to the mixture, we should ask which of the reference library spectra are contained in the given query mixture. This question is inherently asymmetric. Therefore, we should adopt a nonmetric measure. To evaluate our hypothesis, we apply a nonmetric measure formulated by Tversky known as the Contrast Model and compare its performance to the well-known Jaccard similarity index metric on spectroscopic data sets. Our results show that the Tversky similarity measure yields better results than the Jaccard index.Publication A groupware system for virtual product innovation management(Wiley, 2007-10-11) Cormican, Kathryn; O'Sullivan, David; |~|We are experiencing a radical shift in the way organizations are designed, structured, and organized. New organizational forms such as virtual strategic partnerships and networks are replacing traditional bureaucratic, hierarchical organizations. This is particularly evident in the area of product innovation, where organizations are adopting new approaches or ways of working in order to compete. Product innovation management is a complex process because of the range of technical issues that must be addressed and the variety of competencies that must be employed over the life of the development effort. Such initiatives require a substantial investment in terms of resources such as time, money, and effort, all of which are limited. This article focuses on product innovation management in a distributed or virtual environment. It reports on a qualitative case study targeted at product managers. The challenges are identified and discussed. From this analysis, a Web-enabled groupware system is presented.Publication A Framework for Digital Wide-Area Survey from Aerial Photographs(Wordwell Ltd., 1998) Redfern, Sam; |~|Despite the potential offered by modern computer technology to the labour-intensive work of 'Aerial Archaeology', few applications have been published. Where developments have been made, they have tended to be carried out in a somewhat haphazard manner: thereby compounding problems with data standards and consistency: which are essential issues as we move towards a national archaeological database. This paper provides an introduction to digital aerial archaeology and describes the Aerial Archaeology System (AAS) - a computer package proposed as an integral component of the archaeologist's geographical information system (GIS) toolset, which assists the discovery, mapping, and classification of archaeological sites.Publication Collaborative Virtual Environments to Support Communication and Community in Internet-Based Distance Education(jite.org, 2002) Redfern, Sam; Naughton, Niall; |~|In this paper we discuss the use of modern information and communication technologies for distance education (DE) purposes. We argue that current technologies and implementations do not adequately support the key concepts of communication and community that many practitioners believe to be important, particularly if modern pedagogies such as constructivism are to be supported. We propose that collaborative virtual environments (CVEs), which are computer-enabled, distributed virtual spaces or places in which people can meet and interact with others, with agents and with virtual objects, are appropriate tools for improving DE. We discuss the current developments in the areas of CVEs in particular and in computer supported co-operative work (CSCW) in general. We also note those areas in which the majority of CVEs implemented to date have not reached their full potential for DE support, discuss current thought regarding online community, and outline a proposed CVE-based system for DE. The architecture of a CVE should be based on the pedagogical requirements of the community and include three distinct types of virtual space: collaborative zones, common student campus, and lecture rooms. With proper design, a CVE should greatly assist the development of a productive learning community in which students - social, academic, and collaborative needs are metPublication Digital elevation modelling of individual monuments from aerial photographs(John Wiley & Sons, 1999) Redfern, Sam; Lyons, Gerard J.; |~|Digital elevation models (DEM) are crucial data products for a variety of geographic applications, and their generation directly from digitized stereopairs of vertical aerial photographs has recently been accomplished. Despite the growing number of software packages providing softcopy topographic photogrammetry, there is still a need for practical approaches that do not require accurate lens calibration information or time-consuming ground control. This paper presents an algorithm for generating small-area DEMs directly from digitized aerial photographs, with no additional information required other than the flying height of the aircraft and a small number of control points measured from maps.Publication The influence of random interactions and decision heuristics on norm evolution in social networks(Springer, 2011-05) Mungovan, Declan; Howley, Enda; Duggan, Jim; |~|In this paper we explore the effect that random social interactions have on the emergence and evolution of social norms in a simulated population of agents. In our model agents observe the behaviour of others and update their norms based on these observations. An agent's norm is influenced by both their own fixed social network plus a second random network that is composed of a subset of the remaining population. Random interactions are based on a weighted selection algorithm that uses an individual's path distance on the network to determine their chance of meeting a stranger. This means that friends-of-friends are more likely to randomly interact with one another than agents with a higher degree of separation. We then contrast the cases where agents make highest utility based rational decisions about which norm to adopt versus using a Markov Decision process that associates a weight with the best choice. Finally we examine the effect that these random interactions have on the evolution of a more complex social norm as it propagates throughout the population. We discover that increasing the frequency and weighting of random interactions results in higher levels of norm convergence and in a quicker time when agents have the choice between two competing alternatives. This can be attributed to more information passing through the population thereby allowing for quicker convergence. When the norm is allowed to evolve we observe both global consensus formation and group splintering depending on the cognitive agent model used.