# Web-based intelligent distributed and mobile systems The World-wide Web has become a "natural" interaction place for end users, as well as for virtual services and connected objects. The advances of the Web of data, services and Things give way to a new field of research that relies on Semantic Web principles to design high-level, interoperable applications that unify interaction and processing for the different Web actor categories. # Challenges and Contributions ## Intelligent distributed information processing In large unstructured (P2P) networks, there is a growing need for querying and interrelating data structures. We address this challenge by providing a distributed algorithm to maximize mapping discovery between ontologies located on distant peers of the network. This work was developed during the ended <a href="http://liris.cnrs.fr/nicolas.lumineau/projets/simtole/">Simulation pour le Traitement d'Ontologies à Large Echelle (SimTOLE)</a> project. This project led to the development of a simulating environment for semantic resource alignment in P2P network. This environment has evolved to the <a href="http://liris.cnrs.fr/~lmedini/wiki/doku.php?id=simtoleweb">SimTOLEWeb</a> prototype, that allows extracting some peers of the simulator and running them in Web-based clients, locally processing inferences on the peer ontologies. In large information spaces, once resources are aligned and linked, actors have to choose between numerous ones. Recommender systems have been developed to help them make their choices. During <a href="http://liris.cnrs.fr/membres/anciens?id=3502">Simon Meyffret</a>'s PhD thesis, we proposed a local trust-based approach for social recommendation. This approach can be deployed on network peers. It does not require any centralized computation, nor creating additional trust links. However, it behaves well and even outperforms regular centralized approaches in certain conditions. ## Embedded and mobile reasoning For the 21st International World Wide Web conference that was held in Lyon in 2012, we designed and made available the <a href="http://dataconf.liris.cnrs.fr/">DataConf</a> Web application. DataConf is a full client-side Web mashup that allows conference attendees to browse the conference metadata set and enriches these data using other external sources. This application has been reused and evolved to a wider platform: * The <a href="http://www.sympozer.com/">Sympozer</a> comprehensive environment, that allows planifying a conference, provides the corresponding dataset and generates a mobile mashup * The DataConf mashup platform, that embeds a full Web reasoning engine (in JavaScript); this platform is used especially in the recommender engine of the <a href="http://dataconf.liris.cnrs.fr/eswc2013/">ESWC'2013 DataConf instance</a>, that won the 2nd best mashup award of the AI mashup challenge. <div style="width: 100%; text-align: center"> <a href="http://liris.cnrs.fr/actualites/generales/le-liris-et-lapplication-dataconf-obtiennent-la-deuxieme-place-du-ai-mashup-challenge-de-la-conference-eswc2013"><img src="http://liris.cnrs.fr/actualites/generales/le-liris-et-lapplication-dataconf-obtiennent-la-deuxieme-place-du-ai-mashup-challenge-de-la-conference-eswc2013/image" style="max-width: 500px" alt="ESWC award ceremony"></a> </div> The LIRIS transverse <a href="">Web Serviable</a> project federates the DataConf platform and the <a href="http://champin.net/rdfrest/">RDF-REST</a> framework, to help designing Semantic Resource Mashup engines. RDF-REST provides a common HTTP-based semantic interface for RESTful resources, while DataConf provides a mashup engine. Our goal is to issue a full-Web, generic platform that can be reused for diverse applications, among which trace-based systems, service composition engines and Web of Things runtime environments. ## Semantic middleware for Cyber-Physical Systems Cyber-Physical Systems (CPS) provide representations for interlinking physical and virtual worlds. They are often refered to when describing computerized systems that interact with connected objects (sensors, actuators, robots...). In particular, such representations allow leveraging intelligent behaviors for appliances that do not possess sufficient computing resources to develop such behaviors locally, but require the use of network and Web connections to defer such calculations to middleware infrastructures. ### Intelligent adaptation One of our contributions in such infrastructures concerns context-aware adaptation: we use semantic Web principles to ensure interoperability between contextual data, as well as to reason on these data. Contextual reasoning is used both to categorize contextual data into context states and to provide the adaptation mechanisms that select the appropriate services for a given context state. Moreover, storing contextual data in a semantic repository also allows multi-purpose querying of a contextual state, to fulfil different kinds of adaptation goals. This last point is the aim of <a href="http://liris.cnrs.fr/membres?idn=mterdjim">Mehdi Terdjimi</a>'s PhD Thesis. We also used other techniques to perform adaptation and proposed the concept of ideal service variant to diminish computation time in Fuzzy Logics-based adaptation engines (see <a href="http://liris.cnrs.fr/membres/anciens?id=3975">Mounir Beggas</a>'s work). ### Avatar-based architectures for the Web of Things At a global level, intelligent infrastructures aim at reducing the coupling between applications and connected appliances in the Internet of Things. To allow seamless interaction with physical objects though interoperable applications, we chose to rely on Web standards, and therefore, advocate for the <a href="http://www.w3.org/2014/09/wot-ig-charter.html">Web of Things (WoT)</a>. In the <a href="http://liris.cnrs.fr/asawoo">ASAWoO</a> project, we propose a semantic architecture for the WoT, through the notion of avatar. Avatars constitute the "virtual" part of CPS. They allow: * discovering the physical capabilities (sensors, actuators, computing resources) of an object * deducing semantic functionalities from these capabilities * building collaborative functionalities that imply several appliances * deploying applicative code that exploit available functionalities to build user-understandable WoT applications To do this, we rely on other works in different fields, such as Delay-Tolerant Networks to communicate with disconnected objects, efficient cloud infrastructures to minimize energy consumption, multi-level context adaptation to ensure application response time and Multi-Agent Systems to perform inter-avatar collaboration. As standards are an important issue on this very competitive field, we are following (and even sometimes influencing) the current WoT trends. We are involved of the <a href="http://www.w3.org/community/wot/">W3C WoT Interest Group</a> and regularly interact with major actors of the community. Moreover, in our WoT infrastructure, avatars are connected to objects through the <a href="http://liris.cnrs.fr/~lmedini/wiki/doku.php?id=cima">CIMA</a> platform, that is build upon the <a href="http://eclipse.org/om2m/">OM2M</a> platform, following M2M <a href="http://www.etsi.org/">ETSI</a> standards. <div style="width: 100%; text-align: center"> <iframe src="https://www.youtube.com/embed/2SkuJO6s0aM"></iframe> </div> # Contributors * <a href="http://liris.cnrs.fr/membres/anciens?id=31">Frédérique Laforest</a><br/> * <a href="http://liris.cnrs.fr/lionel.medini/">Lionel Médini<br/> * <a href="http://liris.cnrs.fr/membres/anciens?id=3502">Simon Meyffret</a><br/> * <a href="http://liris.cnrs.fr/membres?idn=mterdjim">Mehdi Terdjimi</a><br/> * <a href="http://liris.cnrs.fr/membres/anciens?id=3975">Mounir Beggas</a><br/> ## Grants * Optimacs: ANR PACSLIG project (2009-2012) * <a href="http://liris.cnrs.fr/nicolas.lumineau/projets/simtole/">SimTole</a>: LIRIS transverse project (2009-2010) * <a href="http://liris.cnrs.fr/asawoo/">Adaptive Supervision of Avatars for the Web of Objects (ASAWoO)</a>: ANR INFRA project (2014-2017) * <a href="http://liris.cnrs.fr/asawoo/">Web Serviable</a>: LIRIS transverse project (2014-2015) * <a href="http://liris.cnrs.fr/~lmedini/wiki/doku.php?id=cima">Couche d'Interopérabilité Matériel-Applications (CIMA)</a>: LIRIS platform (2014-2015) # Selected Publications * User-centric Trust-based Recommendation. S. Meyffret, L Médini, F. Laforest. Dans International Conference on Information Technology- New Generations, Shahram Latifi ed. Las Vegas, Nevada, USA. pp. 707-713. Ninth International Conference on Information Technology: New Generations, ITNG 2012, Las Vegas, Nev. ISBN 978-0-7695-4654-4. 2012. * DataConf and Its Linked Open Data Ecosystem: Produce, Link and Consume Scientific Conference Metadata. L Médini, F. Bâcle, F. Le Peutrec, B. Durant de la Pastellière. Dans Proceedings of the LinkedUp Veni Competition on Linked and Open Data for Education held at the Open Knowledge Conference (OKCon 2013), Mathieu d'Aquin, Stefan Dietze, Hendrik Drachsler, Marieke Guy, Eelco Herder ed. Geneva. pp. 3-10. CEUR Workshop Proceedings 1124. ISSN 1613-0076. 2014. <a href="http://ceur-ws.org/Vol-1124/linkedup_veni2013_01.pdf">PDF</a>. * Towards semantic resource mashups. L Médini, P-A. Champin, M. Mrissa, A. Cordier. Dans Services and Applications over Linked APIs and Data (SALAD), workshop at ESWC, Heraklion. pp. 6-9. CEUR Vol-1. 2014. <a href="http://liris.cnrs.fr/Documents/Liris-6685.pdf">PDF</a>. * Ontology Based Context-Aware Adaptation Approach. T Chaari, M. Zouari, F. Laforest. Context-Aware Mobile and Ubiquitous Computing for Enhanced Usability: Adaptive Technologies and Applications. Dragan Stojanovic (Ed) ISBN 978-1-60566-290-9. IGI publishing 2009. * Towards an Ideal Service QoS in Fuzzy Logic-based Adaptation Planning Middleware. M. Beggas, L Médini, F. Laforest, M.T. Laskri. Journal of Systems and Software 92() pp. 71-81, Elsevier, ISSN 0164-1212. 2014. * Semantic Discovery and Invocation of Functionalities for the Web of Things. M. Mrissa, L Médini, J.P. Jamont. Dans IEEE International Conference on Enabling Technologies: Infrastructure for Collaborative Enterprises, Parma. pp. 281-286. 2014. # Software * <a href="http://dataconf.liris.cnrs.fr/">DataConf</a><br/> * <a href="http://www.sympozer.com/">Sympozer</a><br/> * <a href="http://liris.cnrs.fr/~lmedini/wiki/doku.php?id=cima">CIMA</a>