Establishing Semantic Interoperability in a Web of Things (WoT) architecture is a problem that lies at the crossroads of several active research areas. Most connected devices on today’s WoT use heterogeneous data formats. In addition to that, they exchange raw non-semantic data, sometimes in binary format. Web services (which typically serve as the interface between devices and the Web and between users and the Web, they are also responsible for executing complex business logic calculations and accessing databases) cannot efficiently process data in these raw and heterogeneous formats. In addition to that, the advent of the Semantic Web technologies has empowered Web services and firmly established RDF as a universal abstract data model for structured data on the Web, which is why we need to rely on semantically-enabled Web services in order to establish the much needed semantic interoperability between devices.
The first step to achieve this is to « semanticize » the raw input data (we talk about Lifting) in order for Web services to be able to consume it and then « desemanticize » the processing results (we talk about Lowering) to feed them back to the devices. The second step is ensuring the lifted data matches the input description of services consuming it, and ensuring lowered data format is compatible with the device input requirements.
To carry out this second step, we utilize validation. Validation happens on two different levels. First, it happens on the low (non-semantic) level, when devices exchange data with services. Depending on the data format, many validation solutions exist, such as JSON schema and XML schema. Second, it happens on the high (semantic) level, when lifted/lowed data is exchanged with semantically-enabled Web services. We use different technologies to validate this type of data, such as: SHACL, ShEx and SPIN.
The intern will firstly study the state of the art of the different existing RDF graph shape and constraint validation technologies (use the three given examples as a starting point). The study includes comparing complexity, expressivity, flexibility, extensibility, performance and other relevant criteria.
After that, given a set of services and their descriptions together with input data as RDF (a potential use case can be the TERRITOIRE platform services), the intern will have to elaborate a solution to validate the input data against the service input descriptions using the appropriate validation technology (from the initial study).
Alternatively, the intern can propose an elaborate scenario involving a composition of multiple services and their descriptions, where he can apply the appropriate validation technology (from the initial study) in order to determine whether the composition is valid or not, depending on the constraints dictated by descriptions.
This internship is part of the ANR CoSWoT project, also, there is a potential opportunity to pursue a Ph.D. position after the internship.
To apply, send a CV, a motivation letter, grade transcripts, links to public professional profile, references, etc. to: Mahdi BENNARA, firstname.lastname@example.org.