Material Development of the Future

FAIR Data

In a groundbreaking 2016 publication, Mark Wilkinson and Barend Mons introduced the FAIR principles, which have become the foundation for data management in the scientific community. The principles aim to ensure that data is consistently considered findable, accessible, interoperable, and reusable by its providers, users, and owners.

The FAIR Guiding Principles for scientific data management and stewardship Mark D. Wilkinson et al. Scientific Data, 15 March 2016 DOI: 10.1038/sdata.2016.18

The community has been outlined. In this way, the community itself should foster trust and efficiency, from reproducible research to scalable industrial applications and innovation.

Consistently applying the FAIR principles to data management within a community ensures that experiments and simulations can be reproduced under identical conditions in a traceable manner. This strengthens confidence in the results and ensures reliability. Clean, complete datasets improve the quality of machine learning and multiscale models because they enable robust, generalizable predictions.

In materials science and engineering, data from various fields of study are closely interrelated. The properties of the feedstock and the chosen manufacturing route significantly determine the structure and behavior of the final product for a given chemical composition. Accordingly, management of materials data must take into account its diffuse nature.

Accessible and interoperable data accelerates the optimization process in the search for innovative material solutions. Screening processes and the transfer of findings become significantly more efficient. This reduces costly failed attempts and repetitions, saving time and money while making development projects more predictable.

Standardized formats and meaningful metadata facilitate collaboration between teams in companies and scientific institutions. Meanwhile, provenance-oriented documentation, versioning, and metadata ensure traceability, support regulatory requirements, and facilitate audits.

Having consistent data across a component's entire lifecycle allows one to draw valid conclusions about its service life, safety, and sustainability. Because FAIR data is well-documented, it remains usable and valuable in the long term for future analyses and reuse.

In short, managing material data according to FAIR principles increases quality, efficiency, and trust, making it a crucial lever for more successful research and productive implementation in an economic context.

Data-Driven Workflows

What are workflows?

Workflows are structured, repeatable sequences of tasks, activities, or steps that are required to achieve a specific goal or outcome. Businesses frequently use them to increase efficiency, improve communication, and ensure quality results.

Key features of workflows:

• Steps and activities: Workflows consist of several clearly defined steps that can be performed sequentially or in parallel.
• Roles and Responsibilities: Each step can be assigned to specific individuals or groups, who are responsible for carrying it out.
• Automation: Many workflows can be automated to reduce human error and shorten processing time.
• Documentation: Workflows often include documentation that describes how each step should be performed.
• Flexibility: Workflows can be customized or optimized as needed to meet changing requirements.

Examples of workflows:

• Approval Processes: A workflow for approving vacation requests within a company.
• Product development: The steps involved in bringing a new product to market, from brainstorming and design to launch.
• Customer service: The process of handling customer inquiries or complaints.

Workflows are an important tool for process management that help optimize work within organizations.

What are workflows used for?

Companies and organizations use workflows for several reasons:

• Increased efficiency: Workflows standardize and optimize processes, shortening turnaround times and making better use of resources.
• Clarity and structure: They provide a clear framework that makes it easier for employees to understand their tasks and know what to do next.
• Error Reduction: Standardized processes and automation minimize human error, improving the quality of results.
• Transparency: Clear workflows establish responsibilities and provide visibility into the status of tasks, improving communication within the team.
• Traceability: They make it easy to track progress and decisions, which is particularly important for audits and compliance.
• Flexibility: Workflows can be adapted to changing conditions or requirements, keeping organizations agile.
• Knowledge Transfer: Documented workflows facilitate knowledge transfer within a company, especially during the onboarding process for new employees.
• Cost Reduction: Companies can save costs by optimizing processes and reducing errors.
• Improving Employee Satisfaction: Having clear processes and reducing uncertainty can increase job satisfaction because employees can focus on their tasks.
• Customer Satisfaction: Efficient, well-organized workflows lead to faster response times and better service, which increases customer satisfaction.

In summary, workflows play a crucial role in optimizing business processes and contribute to improving the overall performance of an organization.

How do workflows help with digital transformation?

Automation: Workflows play a crucial role in the digital transformation of businesses by linking various processes and technologies. First, they automate recurring tasks, reducing manual intervention and significantly speeding up processes. Automation increases efficiency and reduces the likelihood of errors.

Integration: Another key aspect is the integration of technologies. Workflows lay the foundation for the seamless integration of various digital tools and systems, facilitating data exchange and collaboration between teams. Workflows also support effective data management by helping to capture, store, and organize data. This is crucial for data analysis and subsequent decision-making.

Communication: Digital workflows promote better communication between teams because information and tasks can be shared in real time. This helps ensure that everyone involved is always on the same page. The flexibility and adaptability of these workflows are also important, as they can easily be adapted to new digital requirements, making companies more agile.

Transparency: Another advantage of digital workflows is that they make tasks and processes transparent and traceable, which increases accountability within the team. This is particularly important for management and regulatory compliance. Furthermore, companies can achieve significant cost savings by optimizing their processes and reducing errors, which is of great importance in digital transformation projects.

Scalability: Digital workflows are scalable, allowing companies to adapt as they grow. Their accessibility ensures that information and processes are available anywhere, anytime. This is especially useful for remote work and field operations.

Customer satisfaction: The digitization of workflows improves the customer experience by enabling faster response times and personalized services. Workflows are thus an essential component of digital transformation because they allow companies to work more efficiently, optimize their processes, and adapt to ever-changing digital requirements.

Workflows

SimStack

SimStack (www.simstack.eu) is a commercial workflow environment that enables the efficient design and customization of complex workflows ("rapid prototyping") using drag-and-drop functionality with software modules from various vendors. It displays only the settings relevant to the specific use case. When combined with the automated execution of workflows on mainframes, it minimizes complexity for the end user and reduces the required expertise. It facilitates the transfer of complex scientific multiscale methods to industry.

Pyiron

Pyiron is an integrated development environment (IDE) for computational materials science based on the Python programming language. Developed by the Max Planck Institute for Sustainable Materials (MPI SuSMaT), this workflow environment is continuously refined by PMD and MPI SuSMaT.

The framework's primary objective is to offer a single tool with a unified interface for a variety of simulation, analysis, and visualization tools. With this IDE, users can focus on the science instead of dealing with technical details, such as the input/output formats of the codes and tools.

Workflow Store

The Workflow Store is a platform where scientific work can be shared with others. Researchers can upload workflows and workflow modules to make their research reproducible and accessible to the community.

Data-Driven Ontologies

Understanding the Basics: What Is an Ontology?

An ontology is a structured way of representing knowledge in a specific field. It defines:

1. Concepts and classes: Ontologies describe relevant concepts (or entities) within a specific field of study and organize them into categories or classes.
2. Relationships: They define the connections between concepts, such as hierarchies (e.g., parent-child relationships) and associations ("is a," "has a").
3. Properties: Ontologies also define the properties or attributes associated with concepts and describe their characteristics.
4. Rules and axioms: In many cases, ontologies contain logical rules or axioms that provide more information about concepts and their relationships.
5. Standardization: Ontologies often serve as standards for system interoperability and consistent communication between different applications.

In general, ontologies help organize and structure complex information, thereby improving communication and understanding within a specific field.

Why do we use ontologies?

Ontologies are used in various fields to organize knowledge, promote interoperability between systems, and enable semantic searches. Ontologies improve communication between humans and machines, support the reuse of knowledge and automated processes, and ensure consistency in data usage. Additionally, ontologies facilitate the integration of data from different sources, are extensible, and support decision-making processes. Overall, ontologies help structure complex information and increase efficiency.

By providing shared semantics, ontologies serve as a kind of common dictionary or set of rules that ensures everyone uses the same words with the same meaning when describing materials, components, or processes. Ontologies make data interoperable, meaning that "what you send" is exactly "what I receive."

They create structured, machine-readable data, enabling different systems to communicate with one another. They provide a semantic framework that enables flexible, context-sensitive reasoning. This allows ontologies to handle new or unforeseen user queries that were not explicitly defined in the original data, ensuring their long-term adaptability and relevance.

Areas of Application:

• Knowledge Management: Structuring and organizing knowledge within organizations.
• Artificial intelligence: Improves data processing and machine learning.
• Semantic Web: It enables machines to better understand and process information.

How do ontologies help with digitization?

Ontologies promote digitalization by systematically organizing knowledge and making it easier to store and retrieve information digitally. Ontologies improve the integration of data from various sources, promote semantic interoperability, and optimize communication between digital systems. Ontologies provide a structured knowledge base for AI and data analysis applications, support semantic searches, and increase data reusability across projects. They also enable process automation, improve data quality, and are adaptable. Overall, ontologies create a unified understanding of knowledge, which is essential for digital applications.

Introduction to the Basic Formal Ontology (BFO)

The BFO is a lightweight, domain-neutral upper ontology that serves as a common conceptual foundation for domain-specific ontologies. The BFO provides a small set of philosophically grounded categories and relationships that can be used to consistently model concrete things, processes, and their properties. Examples of these categories and relationships include continuants vs. occurrents and independent vs. dependent entities, qualities, and roles.

What added value does the BFO offer?

• Interoperability: Aligning with the same higher-level structure allows for the easier integration of data and ontologies from different disciplines.
• Consistency and clarity: Clear ontological distinctions (e.g., object versus process) help avoid ambiguity and make models easier to understand.
• Reusability: Domain ontologies that use BFO become more modular and easier to integrate into other projects.
• Support for automatic inference: BFO structures are well-suited for logical checks and reasoners, which allows for the earlier detection of inconsistencies.
• Wide acceptance: Many established ontologies, particularly in the OBO/biomedical domain, are based on BFO, facilitating integration projects.

Ontology Development Kit

The Ontology Development Kit (ODK) is a standardized toolkit and template workflow for developing, validating, publishing, and maintaining ontologies. The ODK bundles proven tools, such as ROBOT, dosdp-tools, Reasoner, and Docker; a Makefile-based build system; CI templates; and automated releases. These features make ontology projects reproducible, maintainable, and team-friendly.

What is the added value of the ODK?

• Standardization: A uniform project structure reduces the effort and errors associated with onboarding.
• Automation: Builds, tests, quality checks, and releases are repeatable and ready for continuous integration (CI).
• Reproducibility: Docker images and Make targets ensure that builds run identically everywhere.
• Interoperability: It supports common formats (OWL/RDF and TTL) and tools (ROBOT and SPARQL).
• Quality Assurance: It performs automatic consistency checks, generates reports, and checks term metadata.
• Scalability: Suitable for individual projects and larger community ontologies.

Getting Started with Protégé: Interactive Work with Ontologies

Protégé is a free and widely used tool for creating and maintaining ontologies (OWL/RDF). It makes domain knowledge explicit, machine-readable, and reusable, which is exactly what is needed to make data interoperable, traceable, and FAIR.

• Protégé formalizes knowledge by clearly defining classes, properties, and relationships so that different teams can share the same semantics.
• It promotes interoperability. Ontologies in OWL/RDF are standard formats that integrate well into data pipelines, Semantic Web services, and machine learning workflows.
• It enables automatic verification. Using reasoners (e.g., HermiT), inconsistencies and logical errors can be detected early on.
• It supports collaboration. WebProtégé offers collaborative editing, commenting, and rights management.
• It is well-established and open, with a large community, many plugins and integrations (OWL API, Jena, and SPARQL), and no licensing costs.
• It increases the value of data. Clear definitions make data easier to find, reuse, and utilize in the long term, which directly contributes to the FAIR principles.

Protégé is a tool that provides semantic clarity, facilitates data integration, and helps build long-term, FAIR-compliant data landscapes. It has relatively low barriers to entry, a strong community, and direct benefits for quality and automation.

Ontologies

PMD Core Ontology

The PMD Core Ontology (PMDco) is a BFO-based mid-level ontology and serves as a semantic anchor in materials science. It is continuously being developed.

The latest version is available on GitHub.

Application ontologies

They describe the specific concepts, relationships, and rules of a particular domain of knowledge or application. Users can describe their own knowledge in an ontology and map it to the PMDco. Published application ontologies are available on GitHub.

OntoDocker

OntoDocker is a Flask web application for managing ontologies. It was developed as part of the MaterialDigital platform. It provides access to ontologies via a graphical user interface (GUI) and an API for management, visualization, querying, and authentication via single sign-on (SSO). OntoDocker uses Docker to deploy ontology databases, such as Blazegraph and Jena Fuseki.

OntoDocker promotes interoperability and data exchange in materials science and engineering.

IT-Architecture

PMD-S

PMD-S instances are PMD servers that interested parties deploy locally. Currently, Pyiron (a workflow environment) and OntoDocker (an ontology management tool) can be added to PMD-S.

These instances are containerized and configured to ensure simple deployment and operation, secure and trustworthy communication, and flexible adaptation to individual requirements.

PMD Mesh

The PMD Mesh is a network of PMD instances that act as Intrusion Detection System (IDS) data providers and consumers. Each instance shares the same foundation: a PMD server (PMD-S). Using Docker and Docker Compose ensures a smooth transition from an initial standalone server to the PMD Mesh. Secure and trustworthy data exchange is based on WireGuard.

Data Portal

The MD DataPortal is a CKAN-based portal for referencing data. Data is stored in a decentralized manner and can be referenced and/or made available via CKAN.

PMD Expert Exchange

The PMD experts are in regular exchange. From relevant event tips to technical details, everything is discussed. We cordially invite you to the following events:

1. PMD Community Interaction
   • Content: General information, networking, project presentations
   • When: Every 2nd Tuesday of the month 1-2 pm
2. Ontology Playground
   • Content: PMDco and application ontology questions, presentations, show & tell sessions
   • When: Bi-weekly Friday 1-2 pm
3. IT Infrastructure
   • Content: Information on the implementation of PMD-S and OntoDocker
   • When: Every last Tuesday of the month 9-10 am
4. Workflow Forum
   • Content: Practical help in developing workflow-related concepts and software tools
   • When: Once a month Wednesday 3:00-4:00 pm
5. Workflows
   • Content: Practical help in developing workflow-related concepts and software tools
   • When: Bi-weekly Friday 8:30-10:00 am