The "10 technology trends" and "5 ways the Internet of Things (IoT) will change your business next year" prognostications are coming out. I always read these and enjoy the critical thinking behind the authors' lists. And every year I think about looking back at last year's lists to see how well the forecasters did, but there isn't much value in "I told you so" whether you are saying it or hearing it so I let that urge pass. I find more value in learning from mistakes and successes in the analysis than judging the forecasts. So this year I thought I would do something a little different.
Data integration is central in Web application development because these applications typically deal with a variety of information formats. Ontology-driven applications face the additional challenge of integrating these multiple formats with the information stored in ontologies. A number of mappings are required to reconcile the variety of formats to produce a coherent overall system. To address these mappings we have developed a number of open source tools that support transformations between some of the common formats encountered when developing an ontology-driven Web application. The Semantic Web Rule Language (SWRL) is a central building block in these tools. We describe these tools and illustrate their use in the development of a prototype Web-based application.
Medical images are proliferating at an explosive pace, similar to other types of data in e-Science. Technological solutions are needed to enable machines to help researchers and physicians access and use these images optimally. While Semantic Web technologies are showing promise in tackling the information challenges in biomedicine, less attention is focused on leveraging similar technologies in imaging. We are developing methods and tools to enable the transparent discovery and use of large distributed collections of medical images in cyberspace as well as within hospital information systems. Our approach is to make the human and machine descriptions of image pixel content machine-accessible through annotation using ontologies.
Most real-world networks are too large to be measured or studied directly and there is substantial interest in estimating global network properties from smaller sub-samples. One of the most important global properties is the number of vertices/nodes in the network. Estimating the number of vertices in a large network is a major challenge in computer science, epidemiology, demography, and intelligence analysis. In this paper we consider a population random graph G = (V;E) from the stochastic block model (SBM) with K communities/blocks. A sample is obtained by randomly choosing a subset W and letting G(W) be the induced subgraph in G of the vertices in W. In addition to G(W), we observe the total degree of each sampled vertex and its block membership. Given this partial information, we propose an efficient PopULation Size Estimation algorithm, called PULSE, that accurately estimates the size of the whole population as well as the size of each community. To support our theoretical analysis, we perform an exhaustive set of experiments to study the effects of sample size, K, and SBM model parameters on the accuracy of the estimates. The experimental results also demonstrate that PULSE significantly outperforms a widely-used method called the network scale-up estimator in a wide variety of scenarios.
Most high-performance expert systems rely primarily on an ability to represent surface knowledge about associations between observable evidence or data, on the one hand, and hypotheses or classifications of interest, on the other. Although the present generation of practical systems shows that this architectural style can be pushed quite far, the limitations of current systems motivate a search for representations that would allow expert systems to move beyond the prevalent "symptom-disease" style. One approach that appears promising is to couple a rule-based or associational system module with some other computational model of the phenomenon or domain of interest. According to this approach, the domain knowledge captured in the second model would be selected to complement the associational knowledge represented in the first module. Simulation models have been especially attractive choices for the complementary representation because of the causal relations embedded in them (Brown & Burton, 1975; Cuena, 1983).