Search

Proc. Conf. of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, We present KELVIN, an automated system for processing a large text corpus and distilling a knowledge base about persons, organizations, and locations. We have tested the KELVIN system on several corpora, including: (a) the TAC KBP 2012 Cold Start corpus which consists of public Web pages from the University of Pennsylvania, and (b) a subset of 26k news articles taken from English Gigaword 5th edition. Our NAACL HLT 2013 demonstration permits a user to interact with a set of searchable HTML pages, which are automatically generated from the knowledge base. Each page contains information analogous to the semi-structured details about an entity that are present in Wikipedia Infoboxes, along with hyperlink citations to supporting text.

IEEE Transactions on Knowledge and Data Engineering (TKDE), Pointwise mutual information (PMI) is a widely used word similarity measure, but it lacks a clear explanation of how it works. We explore how PMI differs from distributional similarity, and we introduce a novel metric, PMImax, that augments PMI with information about a word's number of senses. The coefficients of PMImax are determined empirically by maximizing a utility function based on the performance of automatic thesaurus generation. We show that it outperforms traditional PMI in the application of automatic thesaurus generation and in two word similarity benchmark tasks: human similarity ratings and TOEFL synonym questions. PMImax achieves a correlation coefficient comparable to the best knowledge-based approaches on the Miller-Charles similarity rating dataset.

Second International Workshop on Privacy and Security in Online Social Media (PSOSM), In today’s world, online social media plays a vital role during real world events, especially crisis events. There are both positive and negative effects of social media coverage of events -- it can be used by authorities for effective disaster management or by malicious entities to spread rumors and fake news. The aim of this paper is to highlight the role of Twitter during Hurricane Sandy (2012) to spread fake images about the disaster. We identified 10,350 unique tweets containing fake images that were circulated on Twitter during Hurricane Sandy. We performed a characterization analysis, to understand the temporal, social reputation and influence patterns for the spread of fake images. Eighty six percent of tweets spreading the fake images were retweets, hence very few were original tweets. Our results showed that top thirty users out of 10,215 users (0.3%) resulted in 90% of the retweets of fake images; also network links, such as follower relationships of Twitter, contributed very less (only 11%) to the spread of these fake photos URLs. Next, we used classification models to distinguish fake images from real images of Hurricane Sandy. Best results were obtained from Decision Tree classifier from which we got 97% accuracy in predicting fake images from real. Also, tweet based features were very effective in distinguishing fake images tweets from real, while the performance of user based features was very poor. Our results showed that automated techniques can be used in identifying real images from fake images posted on Twitter.

Second International Workshop on Web of Linked Entities (WoLE), An online user joins multiple social networks in order to enjoy different services. On each joined social network, she creates an identity and constitutes its three major dimensions namely profile, content and connection network. She largely governs her identity formulation on any social network and therefore can manipulate multiple aspects of it. With no global identifier to mark her presence uniquely in the online domain, her online identities remain unlinked, isolated and difficult to search. Literature has proposed identity search methods on the basis of profile attributes, but has left the other identity dimensions e.g. content and network, unexplored. In this work, we introduce two novel identity search algorithms based on content and network attributes and improve on traditional identity search algorithm based on profile attributes of a user. We apply proposed identity search algorithms to find a user's identity on Facebook, given her identity on Twitter. We report that a combination of proposed identity search algorithms found Facebook identity for 39% of Twitter users searched while traditional method based on profile attributes found Facebook identity for only 27.4%. Each proposed identity search algorithm access publicly accessible attributes of a user on any social network. We deploy an identity resolution system, Finding Nemo, which uses proposed identity search methods to nd a Twitter user's identity on Facebook. We conclude that inclusion of more than one identity search algorithm, each exploiting distinct dimensional attributes of an identity, helps in improving the accuracy of an identity resolution process.

AAAI 2013 Fall Symposium on Semantics for Big Data, We describe an approach to reducing the computational cost of identifying coreferent instances in heterogeneous semantic graphs where the underlying ontologies may not be informative or even known. The problem is similar to coreference resolution in unstructured text, where a variety of linguistic clues and contextual information is used to infer entity types and predict coreference. Semantic graphs, whether in RDF or another formalism, are semi-structured data with very different contextual clues and need different approaches to identify potentially coreferent entities. When their ontologies are unknown, inaccessible or semantically trivial, coreference resolution is difficult. For such cases, we can use supervised machine learning to map entity attributes via dictionaries based on properties from an appropriate background knowledge base to predict instance entity types, aiding coreference resolution. We evaluated the approach in experiments on data from Wikipedia, Freebase and Arnetminer and DBpedia as the background knowledge base.

Cold Start task: improvements to our existing KELVIN system, The JHU HLTCOE participated in the Entity Linking and Cold Start Knowledge Base tasks in this year’s Text Analysis Conference Knowledge Base Population evaluation. We have previously participated in TAC-KBP evaluations in 2009, 2010, 2011, and 2012. Our primary focus this year was on the Cold Start task; improvements to our existing KELVIN system included consolidating slot values for an entity, removal of suspect intra-document conference chains, streamlined cross-document entity coreference, and application of inference rules to expand the number of asserted facts.

First International Workshop on Knowledge Extraction and Consolidation from Social Media, Instance consolidation is a way to merge instances that are thought to be the same or closely related that can be used to support coreference resolution and entity linking. For Semantic Web data, consolidating instances can be as simple as relating instances using owl:sameAs, as is the case in linked data, or merging instances that could then be used to populate or enrich a knowledge model. In many applications, systems process data incrementally over time and as new data is processed, the state of the knowledge model changes. Previous consolidations could prove to be incorrect. Consequently, a more abstract representation is needed to support instance consolidation. We describe our current research to perform consolidation that includes temporal support, support to resolve conflicts and an abstract representation of an instance that is the aggregate of a cluster of matched instances. We believe that this model will prove flexible enough to handle sparse instance data and can improve the accuracy of the knowledge model over time.

Proceedings of the 2012 workshop on Data-driven user behavioral modelling and mining from social media, Twitter is a prominent online social media which is used to share information and opinions. Previous research has shown that current real world news topics and events dominate the discussions on Twitter. In this paper, we present a preliminary study to identify and characterize communities from a set of users who post messages on Twitter during crisis events. We present our work in progress by analyzing three major crisis events of 2011 as case studies (Hurricane Irene, Riots in England, and Earthquake in Virginia). Hurricane Irene alone, caused a damage of about 7-10 billion USD and claimed 56 lives. The aim of this paper is to identify the different user communities, and characterize them by the top central users. First, we defined a similarity metric between users based on their links, content posted and meta-data. Second, we applied spectral clustering to obtain communities of users formed during three different crisis events. Third, we evaluated the mechanism to identify top central users using degree centrality; we showed that the top users represent the topics and opinions of all the users in the community with 81% accuracy on an average. The top central people identified represent what the entire community shares. Therefore to understand a community, we need to monitor and analyze only these top users rather than all the users in a community

8th IEEE Int. Conf. on Collaborative Computing: Networking, Applications and Worksharing, Traditional intrusion detection and prevention systems have well known limitations that decrease their utility against many kinds of attacks. Creating a new system that collaboratively combines information from traditional and nontraditional sensors to produce new, relevant signatures is one way to deal with these limitations. In this paper, we present a framework that uses this collaborative approach, as well as the details for a network traffic based classifier that shows promise for detecting malicious traffic.

Social data mining is an interesting phe-nomenon which colligates different sources of social data to extract information. This information can be used in relationship prediction, decision making, pat-tern recognition, social mapping, responsibility distri-bution and many other applications. This paper presents a systematical data mining architecture to mine intellectual knowledge from social data. In this research, we use social networking site facebook as primary data source. We collect different attributes such as about me, comments, wall post and age from facebook as raw data and use advanced data mining approaches to excavate intellectual knowledge. We also analyze our mined knowledge with comparison for possible usages like as human behavior prediction, pattern recognition, job responsibility distribution, decision making and product promoting., 8 pages, 19 figures, 4 tables, 3 equations, ISSN: 2074-904X (Print), ISSN: 2074-9058 (Online)

Proceedings of the Annual Service Research and Innovation Institute Global Conference, Virtualized service models are now emerging and redefining the way information technology is delivered. Managing these services efficiently over the cloud is an open challenge. We are developing a policy-based integrated framework for automating acquisition and consumption of Cloud services. We have developed a tool, Smart Cloud Services tool, to automatically discover, negotiate and consume services from the cloud for a specific use case described by NIST around storage services. This tool incorporates NIST specific definitions for cloud services. The tool makes use of Semantic Web technologies including SPARQL, RDF and OWL to represent and reason about services and service requirements. It is built upon the service lifecycle ontology that we have developed. In this paper, we describe this Smart Cloud Services tool in detail along with the technical challenges we faced in developing it and identify limitations in the current open source cloud computing software.

Cultural Attitudes Towards Technology and Communication 2012, To help us explore parental attitudes towards and usage of technology, we interviewed students ages 10-15 in Hungary and India in focus groups regarding their technology use. This paper focuses on the preliminary results of these students’ perceptions of parental limitations on their technology use. Parents in both countries limited children’s technology use; however, there were differences in the way these limitations were defined and expressed among our participants. The students from Hungary stated that in many situations their parents have a negative attitude toward technology, however, they prescribed fewer rules and gave more freedom to their children to use the technology items discussed. In India, the students indicated that their parents thought technology was useful, a helpful tool, and spent time using technology with their children and in front of their children. Yet the Indian parents limited their children’s use of technology more than the Hungarian parents

We describe a situation-aware intrusion detection system that integrates heterogeneous sources of information to build and maintain a semantically rich knowledge-base about cyber threats and vulnerabilities. Most current intrusion detection and prevention systems rely on signature-based approaches to detect attacks. When an attack signature is not available, such as for a new exploit or a significantly modified known one, such systems are much less effective. Moreover, these intrusion detection systems are point-based solutions which do not make effective use of heterogeneous data sources, which can provide important information related to intrusions which are not yet available as signature patterns. This information can also help detect low-and-slow attacks in which small intrusions that are spatially and temporally apart combine to build a more elaborate attack.

Joint Workshop on Automatic Knowledge Base Construction and Web-scale Knowledge Extraction, Entity linking refers to the task of assigning mentions in documents to their corresponding knowledge base entities. Entity linking is a central step in knowledge base population. Current entity linking systems do not explicitly model the discourse context in which the communication occurs. Nevertheless, the notion of shared context is central to the linguistic theory of pragmatics and plays a crucial role in Grice’s cooperative communication principle. Furthermore, modeling context facilitates joint resolution of entities, an important problem in entity linking yet to be addressed satisfactorily. This paper describes an approach to context-aware entity linking.

Joint Workshop on Automatic Knowledge Base Construction and Web-scale Knowledge Extraction, The steady progress of information extraction systems has been helped by sound methodologies for evaluating their performance in controlled experiments. Annual events like MUC, ACE and TAC have developed evaluation approaches enabling researchers to score and rank their systems relative to reference results. Yet these evaluations have only assessed component technologies needed by a knowledge base population system; none has required the construction of a knowledge base that is then evaluated directly. We describe an approach to the direct evaluation of a knowledge base and an instantiation that will be used in a 2012 TAC Knowledge Base Population track.

Written by the EarthCube Semantic Infrastructure Working Group, In this document, we give a high-level overview of selected Semantic (Web) technologies, methods, and other important considerations, that are relevant for the success of EarthCube. The goal of this initial document is to provide entry points and references for discussions between the Semantic Technologies experts and the domain experts within EarthCube. The selected topics are intended to ground the EarthCube roadmap in the state of the art in semantics research and ontology engineering.

Proceedings of the IEEE Workshop on Semantic Computing and Security, Current state of the art intrusion detection and prevention systems (IDPS) are signature-based systems that detect threats and vulnerabilities by cross-referencing the threat or vulnerability signatures in their databases. These systems are incapable of taking advantage of heterogeneous data sources for analysis of system activities for threat detection. This work presents a situation-aware intrusion detection model that integrates these heterogeneous data sources and build a semantically rich knowledge-base to detect cyber threats/vulnerabilities.

Proceedings of the first International IBM Cloud Academy Conference (ICA CON 2012), Managing and delivering virtualized cloud based services is an open challenge. Current research is focused on specific parts like service discovery; composition etc. and there is no holistic view of what would constitute a lifecycle of virtualized services delivered on a cloud environment. We have developed a policy-based integrated framework for automating acquisition and consumption of Cloud services. This framework divides the cloud service lifecycle into five phases of requirements, discovery, negotiation, composition, and consumption. We have also developed a tool to automatically discover, negotiate and consume services from the cloud for a specific use case described by NIST around storage services. It is built upon the service lifecycle ontology that we have developed. We have built the tool using Semantic Web technologies like SPARQL, RDF and OWL to represent and reason about services and service requirements. This paper describes our methodology and the tool we have developed as well as its implementation on VCL platform.

DUMW Workshop in conjunction with WWW 2012, As the amount of information available on the web has increased, several privacy and security issues around the use of such information have arisen. Government (and private) entities are able to gather and analyze data from several disparate sources with ease. This ability to do large scale analytics of publicly accessible data leads to significant privacy concerns, especially when done by governments. The converse is also true, with concerns about data being shared by individuals and organizations to the web and the cloud. Our work develops a semantically rich, policy driven approach to address the privacy, security and usage concerns around such data.

Proceedings of the 3rd International Workshop on Data Engineering Meets the Semantic Web, Ontology alignment describes a process of mapping ontological concepts, classes and attributes between different ontologies providing a way to achieve interoperability. While there has been considerable research in this area, most approaches that rely upon the alignment of attributes use label-based string comparisons of property names. The ability to process opaque or non-interpreted attribute names is a necessary component of attribute alignment. We describe a new attribute alignment approach to support ontology alignment that uses the density estimation as a means for determining alignment among objects. Using the combination of similarity hashing, Kernel Density Estimation (KDE) and Cross entropy, we are able to show promising F-Measure scores using the standard Ontology Alignment Evaluation Initiative (OAEI) 2011 benchmark.

Users need better ways to explore DBpedia and obtain information from it. Using SPARQL requires not only mastering its syntax and semantics but also understanding the RDF data model, the ontology used by the DBpedia, and URIs for entities of interest. Natural language question answering systems solve the problem, but these are still subjects of research. We describe a compromise in which non-experts specify a graphical “skeleton” for a query and annotate it with freely chosen words, phrases and entity names. The combination reduces ambiguity and allows us to reliably produce an interpretation that can be translated into SPARQL. We demonstrate the approach’s feasibility with an implementation that performs well in an evaluation using queries from the 2011 QALD workshop. Key contributions are the robust methods that combine statistical association and semantic similarity to map user terms to the most appropriate classes and properties used in the DBpedia ontology.

Proc. 5th Text Analysis Conf. (TAC 2012), NIST, Our team from the JHU HLTCOE participated in the Entity Linking and Cold Start Knowledge Base tasks in this year’s Text Analysis Conference Knowledge Base Population evaluation. We have previously participated in TAC-KBP entity linking evaluations in 2009, 2010, and 2011. This year we developed two new systems: CALE (Context Aware Linker of Entities) and KELVIN (Knowledge Extraction, Linking, Validation, and INference) to support our research for this year’s exciting tasks.

Semantic technologies and ontologies play an increasing role in scientific workflow systems and knowledge infrastructures. While ontologies are mostly used for the semantic annotation of metadata, semantic technologies enable searching metadata catalogs beyond simple keywords, with some early evidence of semantics used for data translation. However, the next generation of distributed and interdisciplinary knowledge infrastructures will require capabilities beyond simple subsumption reasoning over subclass relations. In this work, we report from the EarthCube Semantics Community by highlighting which role semantics and ontologies should play in the EarthCube knowledge infrastructure. We target the interested domain scientist and, thus, introduce the value proposition of semantic technologies in a non-technical language. Finally, we commit ourselves to some guiding principles for the successful implementation and application of semantic technologies and ontologies within EarthCube.

Working notes of AAAI Fall Symposium on Open Government Knowledge: AI Opportunities and Challenges, Most open government data is encoded and published in structured tables found in reports, on the Web, and in spreadsheets or databases. Current approaches to generating Semantic Web representations from such data requires human input to create schemas and often results in graphs that do not follow best practices for linked data. Evidence for a table’s meaning can be found in its column headers, cell values, implicit relations between columns, caption and surrounding text but also requires general and domain-specific background knowledge. We describe techniques grounded in graphical models and probabilistic reasoning to infer meaning (semantics) associated with a table using background knowledge from the Linked Open Data cloud. We represent a table’s meaning by mapping columns to classes in an appropriate ontology, linking cell values to literal constants, implied measurements, or entities in the linked data cloud (existing or new) and discovering or and identifying relations between columns.

Research in bioinformatics is a complex phenomenon as it overlaps two knowledge domains, namely, biological and computer sciences. This paper has tried to introduce an efficient data mining approach for classifying proteins into some useful groups by representing them in hierarchy tree structure. There are several techniques used to classify proteins but most of them had few drawbacks on their grouping. Among them the most efficient grouping technique is used by PSIMAP. Even though PSIMAP (Protein Structural Interactome Map) technique was successful to incorporate most of the protein but it fails to classify the scale free property proteins. Our technique overcomes this drawback and successfully maps all the protein in different groups, including the scale free property proteins failed to group by PSIMAP. Our approach selects the six major attributes of protein: a) Structure comparison b) Sequence Comparison c) Connectivity d) Cluster Index e) Interactivity f) Taxonomic to group the protein from the databank by generating a hierarchal tree structure. The proposed approach calculates the degree (probability) of similarity of each protein newly entered in the system against of existing proteins in the system by using probability theorem on each six properties of proteins., 11 pages, 15 figures, 7 tables. arXiv admin note: some text overlap with articles written by other authors, http://bioinformatics.oxfordjournals.org/content/21/15/3234.full , http://www.oxfordjournals.org/nar/database/summary/616 , http://www.jsbi.org/pdfs/journal1/GIW01/GIW01F14.pdf

Proceedings of the Fifth IEEE International Conference on Semantic Computing, Recent years have seen a confluence of two major trends – the increase of mobile devices such as smart phones as the primary access point to networked information and the rise of social media platforms that connect people. Their convergence supports the emergence of a new class of context-aware geosocial networking applications. While existing systems focus mostly on location, our work centers on models for representing and reasoning about a more inclusive and higher-level notion of context, including the user’s location and surroundings, the presence of other people and devices, and the inferred activities in which they are engaged. A key element of our work is the use of collaborative information sharing where devices share and integrate knowledge about their context. This introduces the need for privacy and security mechanisms. We present a framework to provide users with appropriate levels of privacy to protect the personal information their mobile devices are collecting, including the inferences that can be drawn from the information. We use Semantic Web technologies to specify high-level, declarative policies that describe user information sharing preferences. We have built a prototype system that aggregates information from a variety of sensors on the phone, online sources, and sources internal to the campus intranet, and infers the dynamic user context. We show how our policy framework can be effectively used to devise better privacy control mechanisms to control in

Proceedings of the Third IEEE International Conference on Social Computing, Social media services like Twitter, Flickr and YouTube publish high volumes of user generated content as a major event occurs, making them a potential data source for event analysis. The large volume and noisy content of social media makes automatic preprocessing essential. Intuitively, the eventrelated data falls into three major phases: the buildup to the event, the event itself, and the post-event effects and repercussions. We describe an approach to automatically determine when an anticipated event started and ended by analyzing the content of tweets using an SVM classifier and hidden Markov model. We evaluate our performance by predicting event boundaries on Twitter data for a set of events in the domains of sports, weather and social activities.

Proc. Third IEEE International Conference on Information Privacy, Security, Risk and Trust, The balance between privacy and security concerns is a hotly debated topic, especially as government (and private) entities are able to gather and analyze data from several disparate sources with ease. This ability to do large scale analytics of publicly accessible data leads to significant privacy concerns. In particular, for the government, there is the fear of a fishing expedition against individuals. The model in this paper describes a way to address these concerns in a multi-user and multi-database owner environment. The model provides an assurance system where database owners are able to test and audit the assurances given by users thereby increasing the trust in the system. The concept of segregating data used for processing from data needed for final end use and providing different levels of access to them through a mediator machine has been used. The audit component consisting of a justification mechanism increases the trust in the system.

Proceedings of the First International Workshop on Searching and Integrating New Web Data Sources, Vast amounts of information is encoded in structured tables found in documents, on the Web, and in spreadsheets or databases. Integrating or searching over this information benefits from understanding its intended meaning. Evidence for a table's meaning can be found in its column headers, cell values, implicit relations between columns, caption and surrounding text but also requires general and domain-specific background knowledge. We represent a table's meaning by mapping columns to classes in an appropriate ontology, linking cell values to literal constants, implied measurements, or entities in the linked data cloud (existing or new) and discovering or and identifying relations between columns. We describe techniques grounded in graphical models and probabilistic reasoning to infer meaning (semantics) associated with a table. Using background knowledge from the Linked Open Data cloud, we jointly infer the semantics of column headers, table cell values (e.g.,strings and numbers) and relations between columns and represent the inferred meaning as graph of RDF triples. We motivate the value of this approach using tables from the medical domain, discussing some of the challenges presented by these tables and describing techniques to tackle them.

Proceedings of the Second Symposium on Educational Advances in Artificial Intelligence, Intelligent tutoring systems (ITSs) provide students with a one-on-one tutor, allowing them to work at their own pace, and helping them to focus on their weaker areas. The RUR Python Learning Environment (RUR-PLE), a game-like virtual environment to help students learn to program, provides an interface for students to write their own Python code and visualize the code execution (Roberge 2005). RUR-PLE provides a fixed sequence of learning lessons for students to explore. We are extending RUR-PLE to develop the Playing to Program (PtP) ITS, which consists of three components: (1) a Bayesian student model that tracks student competence, (2) a diagnosis module that provides tailored feedback to students, and (3) a problem selection module that guides the student’s learning process. In this paper, we summarize RUR-PLE and the PtP design, and describe an ongoing user study to evaluate the predictive accuracy of our student modeling approach

Proceedings of the AAAI Workshop on Activity Context Representation: Techniques and Languages, We describe work on representing and using a rich notion of context that goes beyond current networking applications focusing mostly on location. Our context model includes location and surroundings, the presence of people and devices, inferred activities and the roles people fill in them. A key element of our work is the use of collaborative information sharing where devices share and integrate knowledge about their context. This introduces a requirement that users can set appropriate levels of privacy to protect the personal information being collected and the inferences that can be drawn from it. We use Semantic Web technologies to model context and to specify high-level, declarative policies specifying information sharing constraints. The policies involve attributes of the subject (i.e., information recipient), target (i.e., the information) and their dynamic context (e.g., are the parties co-present). We discuss our ongoing work on context representation and inference and present a model for protecting and controlling the sharing of private data in context-aware mobile applications.

Proceedings of the Third Canadian Semantic Web Symposium, We describe our ongoing work on using social media as a platform for citizen science. Building on our previous work of facilitating citizen science observations, and using RDF to integrate them with existing biodiversity knowledge, we are currently building Facebook Apps that will enable the reporting of observations, as well as the browsing and tagging of existing observations. The tagging capability serves two main purposes. First, it permits (and, we hope, encourages) multi-stage crowdsourcing for image identification. Second, it serves as a driver of ontology evolution, and permits experiments on potential working relationships between expert-engineered ontologies, and tagbased folksonomies.

Proceedings of the Web Intelligence for Information Security Workshop, The Web is an important source of information about computer security threats, vulnerabilities and cyber-attacks. We present initial work on developing a framework to detect and extract information about vulnerabilities and attacks from Web text. Our prototype system uses Wikitology, a general purpose knowledge base derived from Wikipedia, to extract concepts that describe specific vulnerabilities and attacks, map them to related concepts from DBpedia and generate machine understandable assertions. Such a framework will be useful in adding structure to already existing vulnerability descriptions as well as detecting new ones. We evaluate our approach against vulnerability descriptions from the National Vulnerability Database. Our results suggest that it can be useful in monitoring streams of text from social media or chat rooms to identify potential new attacks and vulnerabilities or to collect data on the spread and volume of existing ones.

Proceedings of the IEEE International Conference on Information Reuse and Integration, In this paper we propose a new form architecture called eXtensible Dynamic Form (XDF) to facilitate data collection process of supplier discovery. This architecture provides users with flexibility to dynamically extend the base form structure with new form components, or to reuse the existing form components by intelligent search. With XDF, the supplier’s capability information can be entered in the structure and format that fit different intentions of individual users. Experimental results demonstrate that the proposed architecture is valuable for facilitating the supplier discovery process.

2011 Symposium on Application Accelerators in High-Performance Computing (SAAHPC'11), Open Computing Language (OpenCL) is fast becoming the standard for heterogeneous parallel com- puting. It is designed to run on CPUs, GPUs, and other accelerator architectures. By implementing a real- world application, a solar radiation model component widely used in climate and weather models, we show that OpenCL multi-threaded programming and execution model can dramatically increase performance even on CPU architectures. Our preliminary investigation indicates that low-level vector instructions and code representations in OpenCL contribute to dramatic performance improvement over the serial version when compared with the execution of the serial code compiled across various compilers on multiple platforms with auto vectorization flags. However, the portability of OpenCL implementations needs to improve, even for CPU architectures.

Proceedings of the IEEE International Symposium on Policies for Distributed Systems and Networks, A Cyber-Physical System (CPS) involves a tight coupling between the physical and computational elements. Security is a key challenge for the deployment of CPS. Therefore, it is highly desirable to extract correct information from a large volume of noisy data and properly evaluate the reputation of reporting devices in CPS. In this paper, we propose a Context-Aware tRust Evaluation scheme for wireless networks in CPS (CARE-CPS), and a set of policy rules are declared to accurately describe how we determine the reputation of each reporting device based on these factors. To validate the CARE-CPS scheme, we have conducted experiments in terms of both simulation and real deployment on smart phones. Experimental results show that the CARE-CPS scheme can properly evaluate the trustworthiness of the report devices in CPS.

12th IEEE International Conference on Mobile Data Management (MDM 2011), Mobile Ad-hoc NETworks (MANETs) are extremely susceptible to various misbehaviors and a variety of trust management schemes have been proposed to detect and mitigate them. Most schemes rely on a set of pre-defined weights to determine how the extent of each misbehavior is used to evaluate the trustworthiness. However, due to the extremely dynamic nature of MANETs, it is not possible to determine a set of weights that are appropriate for all contexts. In this paper, an Automated Trust Management (ATM) system is described for MANETs that uses a support vector machine classifier to detect malicious MANET nodes. The ATM scheme is resilient to attempts by a malicious MANET node to hide its nature by varying its misbehavior patterns over time. The performance of the ATM scheme is evaluated via an extensive simulation study and compared with existing approaches.

Virtualized service models are now emerging and re-defining the way information technology is delivered. Managing these services efficiently over the cloud is an open challenge. In this paper, we describe an integrated methodology for the lifecycle of IT services delivered on the cloud. We have divided the IT service lifecycle into five phases of requirements, discovery, nego-tiation, composition, and consumption. We detail each phase and list the high level ontologies that we have developed for them. We also describe a prototype system that we have developed using Semantic Web technologies to represent and reason about ser-vices and service requirements. This methodology complements previous work on ontologies for service descriptions in that it is focused on supporting negotiation for the particulars of a service and going beyond simple matchmaking.

Proceedings of the 4th International Conference on Pervasive Technologies Related to Assistive Environments, With robotics technology, services can be provided to care for individuals with disabilities. This paper describes an eff ort to improve path planning performance for a robotic arm, resulting in faster user response in real-time. For a robotic arm, particularly with multiple degrees of freedom, path planning is computationally expensive. We propose that it is possible to achieve rapid response times with an assistive robotic arm by caching frequent arm trajectories and creating a "roadmap" of arm movements. By calculating trajectories to possible target goals in advance, we anticipate an improvement in user response times.

Recent years have seen a confluence of two major trends - the increase of mobile devices such as smart phones as the primary access point to networked information and the rise of social media platforms that connect people. Their convergence supports the emergence of a new class of context-aware geo-social networking applications. While existing systems focus mostly on location, our work centers on models for representing and reasoning about a more inclusive and higher-level notion of context, including the user's location and surroundings, the presence of other people and devices, feeds from social networking systems they use, and the inferred activities in which they are engaged. A key element of our work is the use of collaborative information sharing where devices share and integrate knowledge about their context. This introduces the need for privacy and security mechanisms. We present a framework to provide users with appropriate levels of privacy to protect the personal information their mobile devices are collecting including the inferences that can be drawn from the information. We use Semantic Web technologies to specify high-level, declarative policies that describe user's information sharing preferences. We have built a prototype system that aggregates information from a variety of sensors on the phone, online sources, and sources internal to the campus intranet, and infers the dynamic user context. We show how our policy framework can be e ffectively used to devise better privacy control mechanisms to control information ow between users in such dynamic mobile systems.

Proceedings of SRII Global Conference 2011, In Business to Business (B2B) IT outsourcing contracts service providers have to demonstrate value continuously to retain client loyalty. However, sustained value can only be co-created by both service providers and clients though a collaborative process. This process of jointly creating value is termed value co-creation. In this paper we present a new technique that we have devised to identify the opportunities for value co-creation based on priorities for value drivers in an IT enabled B2B services contract. We illustrate this technique by applying it to the IT Helpdesk service.

Proceedings of the Eighth International Workshop on Information Integration on the Web, We present techniques for uncovering links between terror incidents, organizations, and people involved with these incidents. Our methods involve performing shallow NLP tasks to extract entities of interest from documents and using linguistic pattern matching and filtering techniques to assign specific relations to the entities discovered. We also gather more information about these entities from the Linked Open Data Cloud, and further allow human analysts to add intelligent inference rules appropriate to the domain. All this information is integrated in a knowledge base in the form of a graph that maintains the semantics between different types of nodes involved in the graph. This knowledge base can then be queried by the analysts to create actionable intelligence.

Proceedings of the 20th international conference companion on world wide web, This paper presents two-stream processing of audio to index the audio content for Spoken Web search. The first stream indexes the meta-data associated with a particular audio document. The meta-data is usually very sparse, but accurate. This therefore results in a high-precision, low-recall index. The second stream uses a novel language-independent speech recognition to generate text to be indexed. Owing to the multiple languages and the noise in user generated content on the Spoken Web, the speech recognition accuracy of such systems is not high, thus they result in a low-precision, high-recall index. The paper attempts to use these two complementary streams to generate a combined index to increase the precision-recall performance in audio content search. The problem of audio content search is motivated by the real world implication of the Web in developing regions, where due to literacy and affordability issues, people use Spoken Web which consists of interconnected VoiceSites, which have content in audio. The experiments are based on more than 20,000 audio documents spanning over seven live VoiceSites and four different languages. The results suggest significant improvement over a meta-data-only or a speech-recognitiononly system, thus justifying the two-stream processing approach. Audio content search is a growing problem area and this paper wishes to be a first step to solving this at a large scale, across languages, in a Web context.

Proceedings of the SRII Global Conference 2011, Managing the quality of virtualized services that are delivered on the cloud is very challenging. Such services are often composed of smaller components that are assembled on an as-needed basis. In this paper, we propose a framework to measure and semi-automatically track quality delivered by a Virtualized service delivery system. The framework provides a mechanism to relate hard metrics typically measured at the backstage of the delivery process to quality related hard and soft metrics tracked at the front stage where the consumer interacts with the service. This allows administrators responsible for the functioning of a service to monitor its quality based on the measurements typically already done for the component services. The framework is general enough to be applied to any type of IT service. In the paper, we primarily concentrate on the Helpdesk service. We include the performance rules we have created by mining Helpdesk data.

Electronic Business Interoperability: Concepts, Opportunities and Challenges, One of the most critical steps to integrating heterogeneous e-Business applications using different XML schemas is schema mapping, which is known to be costly and error-prone. Past research on schema mapping has not made full use of semantic information imbedded in the hierarchical structure of the XML schema. In this chapter, we investigate the existing schema mapping approaches and propose an innovative semantic similarity analysis approach to facilitate XML schema mapping, merging and reuse. Several key innovations are introduced to better utilize available semantic information. These innovations includes: 1) a layered structure analysis of XML schemas, 2) layer-specific semantic similarity measures, and 3) an efficient semantic similarity analysis using parallel and distributed computing technologies. Experimental results using two different schemas from a real world application demonstrate that the proposed approach is valuable for addressing difficulties in XML schema mapping.

The Proceedings of International Conference on E-business, Management and Economics, We propose a novel solution for semantic-based XML schema matching, taking a mathematical programming approach. This method identifies the globally optimal solution for the problem of matching two XML schemas by reducing the tree-to-tree matching problem to simpler problems of path-topath, node-to-node, and word-to-word matching. We formulate these matching problems as maximum weighted bipartite graph matching problems with different constraints, which are solved by different mathematical programming techniques, including integer programming and dynamic programming. Solutions to simpler problems provide weights for the next stage until the optimal tree-to-tree matching solution is obtained. The effectiveness of this approach has been verified and demonstrated by computer experiments.

Proceeding of the 2010 conference on ECAI 2010: 19th European Conference on Artificial Intelligence, This paper presents a formal convergence proof for EIPFP, an algorithm that integrates low dimensional probabilistic constraints into a Bayesian network (BN) based on the mathematical procedure IPFP. It also extends E-IPFP to deal with constraints that are inconsistent with each other or with the BN structure.

Proceedings of the Poster and Demonstration Session at the 9th International Semantic Web Conference, FOAF is widely used on the Web to describe people, groups and organizations and their properties. Since FOAF does not require unique IDs, it is often unclear when two FOAF instances are co-referent, i.e., denote the same entity in the world. We describe a prototype system that identifies sets of co-referent FOAF instances using logical constraints (e.g., IFPs), strong heuristics (e.g., FOAF agents described in the same file are not co-referent), and a Support Vector Machine (SVM) generated classifier.

Proceedings of the Poster Session of 9th International Semantic Web Conference, Recognizing that two Semantic Web documents or graphs are similar and characterizing their differences is useful in many tasks, including retrieval, updating, version control and knowledge base editing. We describe several text-based similarity metrics that characterize the relation between Semantic Web graphs and evaluate these metrics for three specialc cases of similarity: similarity in classes and properties, similarity disregarding differences in base-URIs, and versioning relation- ship. We apply these techniques for a special use case - generating a delta between versions of a Semantic Web graph. We have evaluated our system on several tasks using a collection of graphs from the archive of the Swoogle Semantic Web search engine.

Proceedings of the Third International Workshop on Social Data on the Web, The friend of a friend (FOAF) vocabulary is widely used on the Web to describe ’agents’ (people, groups and organizations) and their properties. Since FOAF does not require unique ID for agents, it is not clear when two FOAF instances should be linked as co-referent, i.e., denote the entity in the world. One approach is to use logical constraints such as the presence of inverse functional properties as evidence that two individuals are the same. Another applies heuristics based on the string similarity of values of FOAF properties such as name and school as evidence for or against co-reference. Performance is limited, however, by many factors: non-semantic string matching, noise, changes in the world, and the lack of more sophisticated graph analytics. We describe a proto-type system that takes a set of FOAF agents and identifies subsets that are believed to be co-referent. The system uses both logical constraints (e.g., IFPs), strong heuristics (e.g., FOAF agents described in the same file are not co-referent), and an SVM generated classifier. We present initial results using data collected from Swoogle and other sources and describe plans for additional analysis.