Semantic APIs: Give Applications a Common Language

April 04, 2019 by Stefan Summesberger

SEMANTiCS 2019 Proceedings Chair Maria Maleshkova is Assistant Professor at the Computer Science Department III of the University of Bonn. Her research work covers the Web of Things (WoT) and semantics-based data integration topics as well as work in the area of the semantic description of Web APIs, RESTful services and their joint use with Linked Data. In this interview, Maria introduces these areas, provides an overview of the sheer limitless possibilities of future applications and services, and she talks about the great benefits of publications in the SEMANTiCS proceedings.

You are an expert on the Web of Things (WoT) and the semantic description of APIs. What is the WoT and what are the most recent innovations in this area?

The Web of Things aims to achieve a Web, which transparently integrates physical objects. People, buildings, products, cars, machines, and any type of “thing” can be interconnected, exchange data among each other and offer services. In this way the Web can be used as a common platform, which bridges the gap between the physical and digital worlds, paving the pay for new services such as a car, which automatically chooses the correct type of fuel and directly pays for it. In contrast to the Internet of Things (IoT), the WoT adapts already available and widely popular Web protocols and standards, thus ensuring interpretability on the communication level, e.g. all “things” communicate over HTTP.

Currently, the WoT development and evolution are mainly driven by innovative services, which can be provided on top of the integrated devices and the available data. Smart factories, smart homes, spart buildings, smart hospitals and smart assistants are only possible because though the WoT we are able to establish direct communication between previously not connected physical devices and make use of all the gained data. These new services, which bring direct added value, will continue to drive the WoT evolution.

What are the current steps with regards to integrating semantics there? What are the recent breakthroughs and what are the biggest challenges?

The Semantic Web of Things (SWOT) aims to enhance the WoT with semantic technologies. Semantics help to address some of the main challenges in the context of WoT:

  1. providing support for integrating heterogeneous data,

  2. enabling the exchange of heterogeneous data between interfaces, and

  3. facilitating data analytics via reasoning.

Since the main tasks related to adding semantics to the WoT include creating domain models, annotating data and creating RDF datasets, as well as annotating the device interfaces so that they directly provide and consume RDF, I see the main achievements also in these fields: Tools and approaches for annotation, frameworks and architectures for building SWOT systems, and innovative solutions such as a search engine for SWOT.

A domain model is a prerequisite for being able to use semantics and describe the involved devices and data. As a result, the SWOT solutions based on a given domain model, are domain specific. But this hinders the overall cross-domain interoperability and so poses a challenge to developing SWOT systems that are domain independent. Another obvious challenge is dealing with the overhead that is added by the introduction of semantic technologies. Semantics bring a lot of advantages but their use needs to make sense for the specific use case. For smaller scenarios, simple annotations might be more than sufficient, while for more complex systems multiple ontologies and the introduction of rules might be necessary. This trade-off between benefits and overhead is always a challenge when it comes to SWOT.

Do you want to learn more about creating value from your business data? Register NOW for SEMANTiCS 2019!

Tell us about the core benefits when describing APIs semantically? Which new services will become available through this?

Enhancing interfaces in general, and specifically APIs with semantics adds two main benefits. First, this ensures interoperability between the interfaces — the consumed and produced data is in RDF and interface (A) can use the output of interface (B) as its input –  independently of the fact that the proprietary format of the interfaces differs (e.g., XML, JSON or CSV). Second, the search and use of interfaces can be done automatically because each API has a semantic description that specifies its functionality, expected input, resulting output, preconditions, etc. Given the semantic descriptions of interfaces, an engine can automatically search for suitable interfaces, make compositions and create applications that combine the data, not only coming from multiple interfaces but also from different datasets. By enhancing APIs with semantics, these become simply yet another RDF data source.

In terms of services, the very first big success story for semantic APIs was the initial version of Siri. Since then the application has changed a lot but the fundamental idea of easily combining already existing functionalities, such as making reservations, checking the weather and making an appointment, in order to create an innovative service has stayed the same. This is precisely the added value of having semantic APIs – combining or “mashing up” available APIs becomes very easy, since the integration is taken care of, and the variety of services that can be created has no limit.

What are the most exciting research projects that are currently on your radar? Which research results you see implemented in future applications?

Data ownership is becoming more and more important, and this is advocated by new architectural approaches such as the Social Linked Data (Solid) and Industrial Data Space (IDS) ones, which aim to provide solutions that guarantee data ownership. The leading principle behind these data architectures is that the data should never leave the place where it is generated, and any client service should only gain access and link to it, but never copy it. I think that this way of looking at data has the potential to fundamentally change the way that services are developed and provided.

A second development that should be highlighted is in the context of the growing importance of AI. Currently, we are seeing more and more approaches and repositories, which support the reuse and reproducibility of the results of machine learning algorithms. There are collections of implemented neuronal networks, repositories of pre-trained models and training datasets. If this trend is continued, in the near future, we will no longer be working on developing new machine learning approaches but also on combining, configuring and adapting the abundance of already available solutions. With a set of supporting tools anyone would be able to develop AI solutions.

A question to you as the proceedings chair of this year's SEMANTiCS: Scholarly publishing has undergone dramatic changes in the previous years. One of the dominant trends is open access. Please sketch SEMANTiCS' position on this topic?

Open access is very important to us. Since last year SEMANTiCS has a dedicated open access policy. Contributions to the scientific track are published under Gold Open Access LNCS proceedings, for which the conference covers all costs. The poster and demos are published in CEUR proceedings, which are also open access. In addition to that, we are continuously probing additional measures for improving accessibility to the conference outputs and increasing the scientific quality and transparency. Our main priority is to give scientists and experts a platform for their topics and to make high-quality research accessible to the broader public. This is exactly what open access enables and this is why we are in love with open access!

Any last words to those who haven’t submitted yet, researchers facing constraints during the work on their projects or those who still have a lot on their ToDo-Lists before the calls close?

Choose a submission format that suits you and submit! One of the great things about SEMANTiCS is that it offers a range of submission options, from full research papers, to applied research, to short posters and demos. You can choose the best one based on the research work but also based on time constraints and available resources. I always prioritise submitting to SEMANTiCS, since it is a great place to share high-quality research but also one of the best venues in the field for networking, meeting potential collaborators from both industry and research, and getting a sense of what is trendy and hot.

Discuss the potential of Semantic APIs and the Semantic Web of Things with Maria. Register for SEMANTiCS 2019!


The annual SEMANTiCS conference is the meeting place for professionals who make semantic computing work, and understand its benefits and know its limitations. Every year, SEMANTiCS attracts information managers, IT-architects, software engineers, and researchers, from organisations ranging from NPOs, universities, public administrations to the largest companies in the world.