Periodic Reporting for period 2 - OntoTRANS (Ontology driven Open Translation Environment)
Période du rapport: 2021-10-01 au 2023-03-31
The OntoTrans project addresses the need of industry to respond to innovation challenges more efficiently by accessing the relevant information and utilising materials modelling more widely and effectively. In particular, there is a need to strengthen the use of translation as a router supporting end users to get to relevant data and models. To this end, OntoTrans provides a general-purpose ontology-based Open Translation Environment (OTE) able to support the development of dedicated Apps delivering a smart guidance for materials producers and product manufacturers (including associated Translators) through all the steps of the translation process, by:
1. representing manufacturing process challenges in a standard ontological form as technical and business user cases
2. connecting user cases with existing appropriate information sources i.e. available data and materials modelling solutions
3. recommending consistent materials modelling workflow options
4. supporting simulation and validation activities
5. providing semantic results interpretation to facilitate sharing and re-use of user cases and results
The OTE will improve decision making processes via exploratory search and recommendation systems based on an ontology driven knowledge base and its smart integration of knowledge sources, including physics and data-driven models, Open Simulation Platforms (OSP), materials databases and marketplaces. The OTE will be demonstrated at TRL 6 through the in-project development and usage of Apps addressing four selected relevant manufacturing challenges.
Figure 1 Translation steps
The integration of materials modelling and informatics is considered future critical for more agile and sustainable product development and use throughout the entire materials life cycle. It is aligned with the drive for a digital society that assists in developing a circular economy and addresses societal needs. OntoTrans, due to its ontological foundation, delivers deep digitalisation of materials modelling and provides efficient information and knowledge management. It becomes a place for information exchange on key issues concerning materials innovation in industrial applications. Materials and process modelling will be tightly integrated together with experimentation, characterisation, and machine learning (ML) as knowledge sources of emerging digital R&D systems and will therefore contribute to delivering on the promises of AI applications in R&D.
1. representing manufacturing process challenges in a standard ontological form as technical and business user cases
2. connecting user cases with existing appropriate information sources i.e. available data and materials modelling solutions
3. recommending consistent materials modelling workflow options
4. supporting simulation and validation activities
5. providing semantic results interpretation to facilitate sharing and re-use of user cases and results
The OTE will improve decision making processes via exploratory search and recommendation systems based on an ontology driven knowledge base and its smart integration of knowledge sources, including physics and data-driven models, Open Simulation Platforms (OSP), materials databases and marketplaces. The OTE will be demonstrated at TRL 6 through the in-project development and usage of Apps addressing four selected relevant manufacturing challenges.
Figure 1 Translation steps
The integration of materials modelling and informatics is considered future critical for more agile and sustainable product development and use throughout the entire materials life cycle. It is aligned with the drive for a digital society that assists in developing a circular economy and addresses societal needs. OntoTrans, due to its ontological foundation, delivers deep digitalisation of materials modelling and provides efficient information and knowledge management. It becomes a place for information exchange on key issues concerning materials innovation in industrial applications. Materials and process modelling will be tightly integrated together with experimentation, characterisation, and machine learning (ML) as knowledge sources of emerging digital R&D systems and will therefore contribute to delivering on the promises of AI applications in R&D.
During the first and second reporting period, the different components of the OTE have been developed at a foundational level and for the application to the industrial User Cases. The components have been connected, achieving the milestone of a demonstration of the OTE as a whole. Each industrial innovation challenge has been worked on by the respective Translators in depth: the translation without the OTE has been carried out in the first period, followed by translation and execution with support of OTE components.
The OTE software architecture has been collaborative developed and refined, noting the trade-off decisions between key architectural qualities (such as modifiability and ease of development). It provides a detailed view on the OTE components and their functional responsibilities and shows how these map onto the translation process.
Regarding the development specific to OTE components, the following is noted:
(a) Ontology development (EMMO): All four applications have been represented as ontology modules, including representation of objects, processes and relevant properties. EMMO developers further elaborated a middle level ontological layer, common to all applications, aimed at representing innovation challenges in the most seamless way, following the mereotopological and semiotical framework of the EMMO. EMMO foundations for models, workflows, validation and verification, representation of materials etc have been further developed.
(b) Knowledge base (OntoKB): Triple store implementation of the ontology has been carried out including reasoning capabilities. Interfacing and the ability to use different triple story software has been developed. OntoKB is at beta release.
(c) Interfaces to all types of sources of information have been developed up to final release (subject to maintenance), including interfaces to materials databases (via Optimade), simulation platforms, marketplaces and to specific data sources required for the application cases.
(d) OTEAPI, a flexible and modular system for data documentation and dataflow orchestration has been developed. It facilitates seamless information transfer between heterogeneous computer systems, while simultaneously ensuring the information is accurately interpreted and understood.
(e) Exploratory Search System: The ESS has been enhanced with a more solid back-end implementation and integration with OTE tools and with the GUI
(f) Data analytics: initial integration with OTE APIs achieved
(g) Graphical interface, wizard: A first graphical user interface and wizard for executing the application cases has been developed.
Figure 2 OTE Architecture
Each of the aspects of the OntoTrans OTE has been the subject of cross-education and -training for all project partners in this highly interdisciplinary team, for example educating materials modellers on ontologies, triples stores and the knowledge base. Interaction with other projects, in particular Open Innovation Platforms that also utilise ontologies has been very close, facilitated by regular meetings, EMMC Task Groups and joint workshops.
The OTE software architecture has been collaborative developed and refined, noting the trade-off decisions between key architectural qualities (such as modifiability and ease of development). It provides a detailed view on the OTE components and their functional responsibilities and shows how these map onto the translation process.
Regarding the development specific to OTE components, the following is noted:
(a) Ontology development (EMMO): All four applications have been represented as ontology modules, including representation of objects, processes and relevant properties. EMMO developers further elaborated a middle level ontological layer, common to all applications, aimed at representing innovation challenges in the most seamless way, following the mereotopological and semiotical framework of the EMMO. EMMO foundations for models, workflows, validation and verification, representation of materials etc have been further developed.
(b) Knowledge base (OntoKB): Triple store implementation of the ontology has been carried out including reasoning capabilities. Interfacing and the ability to use different triple story software has been developed. OntoKB is at beta release.
(c) Interfaces to all types of sources of information have been developed up to final release (subject to maintenance), including interfaces to materials databases (via Optimade), simulation platforms, marketplaces and to specific data sources required for the application cases.
(d) OTEAPI, a flexible and modular system for data documentation and dataflow orchestration has been developed. It facilitates seamless information transfer between heterogeneous computer systems, while simultaneously ensuring the information is accurately interpreted and understood.
(e) Exploratory Search System: The ESS has been enhanced with a more solid back-end implementation and integration with OTE tools and with the GUI
(f) Data analytics: initial integration with OTE APIs achieved
(g) Graphical interface, wizard: A first graphical user interface and wizard for executing the application cases has been developed.
Figure 2 OTE Architecture
Each of the aspects of the OntoTrans OTE has been the subject of cross-education and -training for all project partners in this highly interdisciplinary team, for example educating materials modellers on ontologies, triples stores and the knowledge base. Interaction with other projects, in particular Open Innovation Platforms that also utilise ontologies has been very close, facilitated by regular meetings, EMMC Task Groups and joint workshops.
The state of the art of data re-use and industry commons is that despite efforts to create widely accessible data repositories, there is a limited reuse due to a lack of knowledge about the data. OntoTrans contributes strongly to overcoming this issue by establishing widely agreed ontologies and taxonomies.
The project ambition is to be a blueprint for R&D digitalisation of research intensive, high-value-add industries in Europe. Key to that is a strong semantic basis. OntoTrans will progress beyond the state-of-the-art on the following expected results:
- ontology for chemicals, materials, properties, modelling, processing and manufacturing
- knowledge base systems and recommendation system
- interfaces providing flexible data documentation and dataflow orchestration
- robust exploratory search system with connection to ontologies and ready for industrial applications
- digitalisation of manufacturing challenge applications
OntoTrans OTE provides industry an effective means to widely benefit from a model-based approach for a better understanding of processing, properties and characterisation from chemistry to application performance. This will subsequently cut cost and time by avoiding unnecessary experiments, better targeting and reduced usage of expensive and dangerous materials. Furthermore, the ability to invest resources in a more targeted and effective way will be increased. OntoTrans is synergistic with Big Data and industry digitalisation, leading to smarter decisions and efficient use of resources.
The project ambition is to be a blueprint for R&D digitalisation of research intensive, high-value-add industries in Europe. Key to that is a strong semantic basis. OntoTrans will progress beyond the state-of-the-art on the following expected results:
- ontology for chemicals, materials, properties, modelling, processing and manufacturing
- knowledge base systems and recommendation system
- interfaces providing flexible data documentation and dataflow orchestration
- robust exploratory search system with connection to ontologies and ready for industrial applications
- digitalisation of manufacturing challenge applications
OntoTrans OTE provides industry an effective means to widely benefit from a model-based approach for a better understanding of processing, properties and characterisation from chemistry to application performance. This will subsequently cut cost and time by avoiding unnecessary experiments, better targeting and reduced usage of expensive and dangerous materials. Furthermore, the ability to invest resources in a more targeted and effective way will be increased. OntoTrans is synergistic with Big Data and industry digitalisation, leading to smarter decisions and efficient use of resources.