A Polyglot Domain Model Library: narrowing the gap between Semantics Engineers, Domain and Standardization Experts, and Software Developers
A Talk by Veronika Haderlein-Høgberg PhD (Senior Researcher, Semantics & Reasoning, Siemens Foundational Technologies, Data & AI, Siemens AG, Siemens AG)
About this Talk
Developing and maintaining Domain Models of various kinds – both formal OWL-based ones as well as models based on a proprietary formalism - has become a well-established process in many parts of the manufacturing industry.
However, the uptake of these Domain Models among Software Developers – in the sense of a structured approach to letting these models formally inform software code and interfaces - is slow.
This results in the data which software systems produce, not being readily available for querying or reasoning in Knowledge Graph applications, and thus leads to an extensive need for complicated mappings and – on an organizational level - a continued hesitation towards the adoption of Knowledge Graphs.
Equally, there is regularly a disconnect between Domain Experts developing and maintaining standards and norms about their field of knowledge on the one hand, and the expression of that knowledge in formal ontologies, leading to the fact that industry standards which typically are written in natural language, i.e. in linear text, or proprietary formalisms, need to be translated by a Semantics Engineer into a formal ontology in order to make them available to Knowledge Graph-applications.
These processes are notoriously slow and tedious, as Semantics Engineers typically lack the in depth-knowledge of a given domain that is necessary to cover it correctly and in every detail in an ontology.
Finally, the manufacturing industry is characterized by a large heterogeneity of standards and norms that govern the way data is expressed, exchanged, and reviewed for compliance, further stressing the necessity of turning the process of representing standards and norms as formal Domain Models into a centrally available, streamlined service.
Assuming that each of the involved groups of professionals act as they do out of reasons that make sense in the context of the constraints of their work – the need for ready-made code libraries and data contracts that are easy to implement; the need to focus on the domain rather on the grammar of formal semantics; the need for fully expressive ontologies in order to support the rich potential of Knowledge Graphs – we are currently developing an approach that intends to narrow these gaps in that it offers the expression of a Domain Model along multiple formalisms, so that one and the same domain can manifest itself as full-length Industry Standard-document or an Open World-based ontology in OWL, but can also be represented as carefully curated templates for semantically enabled data contracts along industry data standards such as OPC-UA, AAS or proprietary JSON-schemas.
We call this a “Polyglot Domain Model Library” in that it “speaks” different formal languages. It revolves around a set of on-demand-translation services in between various data contract formalisms, supporting tools to query and explore Standard documents to ease the work of Semantics Engineers, and LLM-based approaches to support the decision making that is required for going from an Open World-ontology to a Closed World-data contract.
We observe the slow adoption of Knowledge Graphs due to the above-mentioned gaps in many parts of the manufacturing industry and believe that our approach of a Polyglot Domain Model Library will help lowering the adoption threshold, especially in large, heterogeneous organizations.
Target audience: Software Architects, Data Governance Professionals, Ontology Engineers, Standardization Experts
About The Speakers
Senior Researcher, Semantics & Reasoning, Siemens Foundational Technologies, Data & AI, Siemens AG, Siemens AG
Dr. Veronika Haderlein-Høgberg is a Senior Researcher in the field of Knowledge Modelling, with many years of experience from various industries (Offshore, Defence, Public Sector, Manufacturing), having given her in-depth insights into how to best bring different communities of practice together.
We’re a cloud tech company that provides organisations around the world with computing infrastructure and software to help them innovate, unlock efficiencies and become more effective. We also created the world’s first – and only – autonomous database to help organise and secure our customers’ data.
