Anwar Vahed

This video contains the recording of the presentation titled "Data-Intensive Research Infrastructure for South Africa (DIRISA)" given at the Sci-GaIA User Forum & Final Conference, held on the 23th and the 24th of March 2017 at the CSIR International Convention Centre (CICC), Pretoria, South Africa. For more information about the event, please visit http://agenda.ct.infn.it/e/scigaia-final-event. For more information about the Sci-GaIA project, please visit http://www.sci-gaia.eu

Virtual observatories mature from their original domain and become common practice for earth observation research and policy building. The term Virtual Observatory originally came from the astronomical research community. Here, virtual observatories provide universal access to the available astronomical data archives of space and ground-based observatories. Further on, as those virtual observatories aim at integrating heterogeneous ressources provided by a number of participating organizations, the virtual observatory acts as a coordinating entity that strives for common data analysis techniques and tools based on common standards. The Sensor Web is on its way to become one of the major virtual observatories outside of the astronomical research community. Like the original observatory that consists of a number of telescopes, each observing a specific part of the wave spectrum and with a collection of astronomical instruments, the Sensor Web provides a multi-eyes perspective on the current, past, as well as future situation of our planet and its surrounding spheres. The current view of the Sensor Web is that of a single worldwide collaborative, coherent, consistent and consolidated sensor data collection, fusion and distribution system. The Sensor Web can perform as an extensive monitoring and sensing system that provides timely, comprehensive, continuous and multi-mode observations. This technology is key to monitoring and understanding our natural environment, including key areas such as climate change, biodiversity, or natural disasters on local, regional, and global scales. The Sensor Web concept has been well established with ongoing global research and deployment of Sensor Web middleware and standards and represents the foundation layer of systems like the Global Earth Observation System of Systems (GEOSS). The Sensor Web consists of a huge variety of physical and virtual sensors as well as observational data, made available on the Internet at standardized interfaces. All data sets and sensor communication follow well-defined abstract models and corresponding encodings, mostly developed by the OGC Sensor Web Enablement initiative. Scientific progress is currently accelerated by an emerging new concept called scientific workflows, which organize and manage complex distributed computations. A scientific workflow represents and records the highly complex processes that a domain scientist typically would follow in exploration, discovery and ultimately, transformation of raw data to publishable results. The challenge is now to integrate the benefits of scientific workflows with those provided by the Sensor Web in order to leverage all resources for scientific exploration, problem solving, and knowledge generation. Scientific workflows for the Sensor Web represent the next evolutionary step towards efficient, powerful, and flexible earth observation frameworks and platforms. Those platforms support the entire process from capturing data, sharing and integrating, to requesting additional observations. Multiple sites and organizations will participate on single platforms and scientists from different countries and organizations interact and contribute to large-scale research projects. Simultaneously, the data- and information overload becomes manageable, as multiple layers of abstraction will free scientists to deal with underlying data-, processing or storage peculiarities. The vision are automated investigation and discovery mechanisms that allow scientists to pose queries to the system, which in turn would identify potentially related resources, schedules processing tasks and assembles all parts in workflows that may satisfy the query.

Current research in modular neural networks (MNNs) have essentially two aims; to model systematic methods for constructing neural networks of high complexity and secondly, to provide building blocks for hybrid symbolic- and connectionist knowledge based implementations. The principal benefit of MNNs is that it combines the desirable features of different neural network architectures while compensating for their individual weaknesses. This paper reviews several models of modular neural networks and describes a method for constructing modular neural network subroutines that facilitate easier knowledge extraction. We explore this feature and further consider the generalization abilities of network subroutines as compared with conventional neural network architectures

This video belongs to the playlist of all the recordings related to the Sci-GaIA User Forum & Final Conference, held on the 23th and the 24th of March 2017 at the CSIR International Convention Centre (CICC), Pretoria, South Africa. For more information about the event, please visit http://agenda.ct.infn.it/e/scigaia-final-event For more information about the Sci-GaIA project, please visit http://www.sci-gaia.eu

Even though open data has a number of benefits, the biggest challenge facing open data is privacy. Opening access to such data involves trading off privacy for utility or vice versa. Releasing the raw data allows for better engagement with the data; however, this creates privacy risks. Protecting the data limits the usefulness of the data. Therefore a balance between privacy and utility must be maintained. It therefore becomes challenging to release data while ensuring that it is useful. This paper aims to understand the limits of open data in terms of privacy. Given national regulatory requirements in South Africa to ensure data privacy such as the Protection of Personal Information (POPI) Act, appropriate measures for mitigating privacy violation are particularly relevant for the Data Intensive Research Initiative of South Africa (DIRISA), a government funded enterprise to manage and provide access to research data repositories. This paper investigates the currently existing South...

In numerous research elds, scientists are increasingly turning to e-science and specifically scienti c work ows as a way of improving, broadening and hastening their results. Enhanced collaboration, on-demand access to tools, data and high performance processing facilities are some of the gains to be made. Scienti c work ows are concerned with, amongst others, supporting the repeatability and provenance of experiments. Scienti c work ows have had a measure of success in the astronomy, bio-informatics, chem-informatics, geophysics and eco-informatics domains [Gil, et. al., 2007]. This paper describes our initial investigations into developing geospatial Scienti c Work ows to support researchers in exploring, integrating and visualising Earth Observation and GIS data in conjunction with other research data. We describe some of the functionalities we require in the context of three sets of research endeavour- wild re research, ood modelling and the linking of disease outbreaks to multi...

Abstract. Statistics South Africa (Stats SA) has identified the centralised storage of data and metadata as a key element in improving the quality of data. This has been interpreted within the Data Management and Information Delivery (DMID) project of Stats SA, to entail a solution founded on two key technological deliverables: There would be a central metadata repository where everything required to interpret and understand the data is stored according to standard, uniform and agreed fields and formats. Secondly, a central data storage facility would consolidate disparate sources of statistical data in a standard manner, together with a set of tools for retrieval, analysis and report-generating data. Additional core enablers comprise of the identification, development and adoption of standards, and approved policies and procedures regulating data management. This factorisation of the DMID project is reflected in its current organisational structure. This paper commences with our no...

Climate change poses a major threat to environment al sustainability. Africa in particular, is vulnerable with projected worsening food security, increased threats to public health, increased stress on surfa ce water resources and a general increase in extreme weather events. Political commi tment regarding the development of effective mitigation and adaptation strategies n eeds to be complemented by a more comprehensive understanding of the effects of climate change. However, comparatively little is known about the projected e ffects of climate change for Africa at the local scale. We present data sets of climate -change projections downscaled to a resolution of 0.5° over the African continent, and data sets downscaled further to an 8 km grid size for areas covering selected urban ce ntres in Africa. These data sets constitute a novel and unique resource for research . We describe a set of web services that provide access to aggregations of sel ected key climate-change variables de...

Climate change poses a major threat to environmental sustainability. In Africa, sectors which face projected vulnerability include, food security, public health and water resources. A general increase in extreme weather events is also predicted and the effects thereof are beginning to be felt. In order for responsible parties and stake holders to respond to these threats, research needs to be undertaken to understand the effects of climate change. Consequently, data has to be available for such research. The rapid advance in computer technology has led to heavy reliance on digital predictive modelling and digital data dissemination which have both been leveraged in this project. We present data sets of predicted climate-change variables, over selected urban regions of the African continent from six different types of models, which constitute a novel research resource. These datasets can be used in a host of further studies on how to address effects of climate change and can help sta...

The paper focuses on the information technology infrastructure required for the evaluation and monitoring of risk relating to floods in South Africa. It may be argued that in the context of developing countries, flood preparedness is more valuable than the actual response to a flood disaster. The paper looks at this flood preparedness in the context of informal and semi-formal settlements. An information technology infrastructure is proposed that will allow decision makers to be alerted to possible flood high risk areas, and in so doing maximise preparedness.

ABSTRACT Information Society Technologies (IST) Africa 2009, Uganda, 6-8 May, 2009 Despite many international and local initiatives on disaster risk management and advances in scientific knowledge, the social and economic impact of natural disasters in emerging and developing countries is still increasing. Various activities are under way to investigate the potential of counter measures and mitigation strategies to handle the growing number of natural disasters. In June 2008, the European Commission initiated a new research project to demonstrate the capacity of standardised low cost interoperable information and communication technology (ICT) solutions to effectively mitigate disaster risk by addressing all phases of disaster risk management from risk assessment to recovery; paving the way to improved risk governance and contributing to sustainable development. This paper will present the first results from a South African perspective, which provides an interesting insight on the major challenges ahead, where technological progress meets organisational reality. It illustrates the current situation in South Africa in the context of governmental instruments and organisation to support disaster risk reduction and disaster management and sheds light on our recent achievements in scientific workflows for disaster management research. A flooding scenario is used to demonstrate the functionality of scientific workflows. The paper concludes with an outlook how the entire risk management environment can benefit from them.

Statistics South Africa (Stats SA) has identified the centralised storage of data and metadata as a key element in improving the quality of data. This has been interpreted within the Data Management and Information Delivery (DMID) project of Stats SA, to entail a solution founded on two key technological deliverables: There would be a central metadata repository where everything required to interpret and understand the data is stored according to standard, uniform and agreed fields and formats. Secondly, a central data storage facility would consolidate disparate sources of statistical data in a standard manner, together with a set of tools for retrieval, analysis and report-generating data. Additional core enablers comprise of the identification, development and adoption of standards, and approved policies and procedures regulating data management. This factorisation of the DMID project is reflected in its current organisational structure. This paper commences with our notions of d...

Scrum provides many benefits to organizations requiring a project management framework for complex adaptive problems. Some of these benefits include improved teamwork, improved time to market, and a noticeable decrease in software defects. The primary objective of this paper is to test nineteen research hypotheses that require a quantitative analysis of the Scrum framework. In order to test these hypotheses, the findings of a survey questionnaire was used to gather response data from Scrum practitioners on their perceptions of factors affecting Scrum adoption. Exploratory factor analysis and Cronbach’s alpha analysis confirmed the validity and reliability of the measuring instrument. Following these analyses, a correlation matrix was used to test the relationship strength among the different factors. The Spearman correlation analysis revealed statistically significant correlations. Multiple linear regression statistical models were developed to examine the existence of factors and c...

Scrum is the most adopted Agile methodology. The research conducted on Scrum adoption is mainly qualitative and there is therefore a need for a quantitative study on Scrum adoption challenges. The primary objective of this paper is to present the findings of a study on the factors that have a significant relationship with Scrum adoption as perceived by Scrum practitioners working within South African organizations. Towards this objective, a narrative review to extract and synthesize the existing challenges was conducted. These synthesized challenges were used in the development of a conceptual framework for evaluating the challenges that have a correlation and linear relationship with Scrum adoption. Following this, a survey questionnaire was used to test and evaluate the factors forming part of the developed framework. The findings indicate that Relative Advantage, Complexity, and Sprint Management are factors that have a significant linear relationship with Scrum adoption. Our rec...

... 5 Machine Learning Method ... Complex rnodcls tend to overfit the given data, and thus show poorgeneralization pcrforrnance cornpaxed to simpler DFAs which ex-plain thc ... Figure 3: Prefix Tree: Shown is a prefix tree of depth (ie string length) 3 for the random 10-Statc DFA ...

CLUVA (CLimate change and Urban Vulnerability in Africa; http://www.cluva.eu/) is a 3 years project, funded by the European Commission in 2010. Its main objective is the estimate of the impacts of climate changes in the next 40 years at urban scale in Africa. The mission of CLUVA is to develop methods and knowledge to assess risks cascading from climate-changes. It downscales IPCC climate projections to evaluate threats to selected African test cities; mainly floods, sea-level rise, droughts, heat waves and desertification. The project evaluates and links: social vulnerability; vulnerability of in-town ecosystems and urban-rural interfaces; vulnerability of urban built environment and lifelines; and related institutional and governance dimensions of adaptation. A multi-scale and multi-disciplinary quantitative, probabilistic, modelling is applied. CLUVA brings together climate experts, risk management experts, urban planners and social scientists with their African counterparts in a...

Virtual observatories mature from their original domain and become common practice for earth observation research and policy building. The term Virtual Observatory originally came from the astronomical research community. Here, virtual observatories provide universal access to the available astronomical data archives of space and ground-based observatories. Further on, as those virtual observatories aim at integrating heterogeneous ressources provided by a number of participating organizations, the virtual observatory acts as a coordinating entity that strives for common data analysis techniques and tools based on common standards. The Sensor Web is on its way to become one of the major virtual observatories outside of the astronomical research community. Like the original observatory that consists of a number of telescopes, each observing a specific part of the wave spectrum and with a collection of astronomical instruments, the Sensor Web provides a multi-eyes perspective on the c...

Neural networks do not readily provide an explanation of the knowledge stored in their weights as part of their information processing. Until recently, neural networks were considered to be black boxes, with the knowledge stored in their weights not readily accessible. Since then, research has resulted in a number of algorithms for extracting knowledge in symbolic form from trained neural networks. This article addresses the extraction of knowledge in symbolic form from recurrent neural networks trained to behave like deterministic finite-state automata (DFAs). To date, methods used to extract knowledge from such networks have relied on the hypothesis that networks' states tend to cluster and that clusters of network states correspond to DFA states. The computational complexity of such a cluster analysis has led to heuristics that either limit the number of clusters that may form during training or limit the exploration of the space of hidden recurrent state neurons. These limit...

... 5 Machine Learning Method ... Complex rnodcls tend to overfit the given data, and thus show poorgeneralization pcrforrnance cornpaxed to simpler DFAs which ex-plain thc ... Figure 3: Prefix Tree: Shown is a prefix tree of depth (ie string length) 3 for the random 10-Statc DFA ...

Scientific Workflows provides a technology that facilitates researchers by allowing them to capture in a machine processable manner the method relating to some research. This increases both provenance and repeatability of the research and allows for increased collaboration through workflow sharing. The Sensor Web is an open complex adaptive system the pervades the internet and provides access to sensor resources.

ABSTRACT Information Society Technologies (IST) Africa 2009, Uganda, 6-8 May, 2009 Despite many international and local initiatives on disaster risk management and advances in scientific knowledge, the social and economic impact of natural disasters in emerging and developing countries is still increasing. Various activities are under way to investigate the potential of counter measures and mitigation strategies to handle the growing number of natural disasters. In June 2008, the European Commission initiated a new research project to demonstrate the capacity of standardised low cost interoperable information and communication technology (ICT) solutions to effectively mitigate disaster risk by addressing all phases of disaster risk management from risk assessment to recovery; paving the way to improved risk governance and contributing to sustainable development. This paper will present the first results from a South African perspective, which provides an interesting insight on the major challenges ahead, where technological progress meets organisational reality. It illustrates the current situation in South Africa in the context of governmental instruments and organisation to support disaster risk reduction and disaster management and sheds light on our recent achievements in scientific workflows for disaster management research. A flooding scenario is used to demonstrate the functionality of scientific workflows. The paper concludes with an outlook how the entire risk management environment can benefit from them.

... Distributed computing architecture for SW4SW:the OGC Sensor Web Enablement (SWE) [10] initiative provides an open distributed computing infrastructure where sensor resources can be published, automatically discovered and accessed over the web. ...