Following our earlier blog post on ‘Determining Compatbility’, we have teamed up with our partners at metaphacts to create an innovative knowledge graph-based application. The application is built on top of metaphactory, a knowledge graph management, visualisation and interaction platform, and RDFox, our knowledge graph and semantic reasoning engine.
Together, metaphactory and RDFox deliver unprecedented results in compatibility determination scenarios by allowing users to quickly and efficiently gain access to actionable and meaningful insights. This blog will demonstrate the functionality of the metaphactory & RDFox joint solution, using an industrial configuration use case example.
This blog was co-written by Irina Schmidt (metaphacts).
Thank you to Ilija Kocev, Daniel Herzig (metaphacts) and Valerio Cocchi (Oxford Semantic Technologies) for their work on the demo system.
Knowledge graphs have structural advantages over traditional models for solving compatibility solutions. They overcome the flexibility limitations of relational databases as data is stored as richly connected entities, the system can be updated easily, and end-users can undertake targeted search, discovery and exploration. Knowledge graphs ensure data is FAIR — Findable, Accessible, Interoperable and Reusable.
Determining compatibility is integral to many business models and processes. The process requires the user to check millions of combinations, to assess whether components fit together, or if components meet specified requirements. Additional factors may also need to be taken into account, for example, regulations or customer budgets.
With our knowledge graph-driven application for industrial configuration management, support engineers, product managers, technical planners, or technical maintenance specialists can evaluate components such as motors, gears, switches, power supplies and controllers, and their individual characteristics. The user can determine how components fit together and how they can be combined to create solutions that solve very specific customer needs or maintenance requests, while staying within a predefined budget.
Let’s consider the example of a support engineer who is looking for a complete rotation solution for one of her customers. The solution should come with a particular brushless motor, a specific power supply, a minimum speed (50 rpm) and torque (500 Nm), but should stay within a certain budget (max. 129€).
Compatibility between the components represented in the knowledge graph is determined by reasoning (datalog rules) applied by RDFox. RDFox also automatically computes the cost of each solution. Using an intuitive interface built with metaphactory, the engineer can explore available solutions by simply selecting the components that should be included and defining the constraints that should be respected.
Once she has found a few solutions that satisfy her needs, the engineer can proceed to look at each one in detail and decide on the one she wants to order for her customer.
Now let’s say our support engineer is assigned to replace a faulty DC motor at a customer site, but all other components that are part of the customer’s existing rotation solution should be kept. Using the configuration management application built with metaphactory and RDFox, she can start by looking at the various DC motors available and filter down to find the one compatible with the technical setup the customer has in place.
For example, she might be looking for a DC motor with a minimum torque of 200 Nm and a provided speed between 2,000 and 4,000 rpm, but which should not go over a predefined budget of €20.
After a quick search based on her parameters, our engineer can go on to further explore her search results, for example which other components the resulting DC motors are compatible with:
…and which rotation solutions these DC motors are part of:
She can also look at each DC motor’s specifications in detail:
…or visually explore relationships between a DC motor and other components:
Using the visual exploration component that is integrated into the application, our end user can quickly build a diagram to show power supplies this DC motor is compatible with, as well as rotation solutions it is part of. The engineer can also find out at a glance which information was initially loaded into RDFox (e.g., the grey “type” relationship tells us that this information is core information) and which information was inferred based on the rules defined with RDFox (e.g., the red “compatibleWith” or “component” relationships tell us that this is inferred information).
With its keyword interpretation engine and its intuitive visual graph exploration component, metaphactory allows end users to leverage core and inferred graph data in RDFox and perform targeted, natural-language queries that deliver instant results and can be explored further to discover previously unknown connections. In the example below, our engineer is searching power supplies compatible with a particular DC motor (DCMotor12) to set up a rotation solution for a customer. She starts by defining a customer budget of €80 for the complete solution.
Then, she searches for all power supplies compatible with DCMotor12. She can type her keywords in the search bar and the system immediately returns a visual graph with all matching components. Note how metaphactory’s keyword interpretation engine understood that the keyword “powersupply” refers to the “DCPowerSupply” entities stored in the knowledge graph.
She then explores the data in the system further to find the rotation solutions that include these components. Note that two rotation solutions including two different power supplies are within budget, while a third one slightly exceeds the budget.
From here, our engineer can further explore relations in the graph and discover that two of the power supplies listed are also compatible with a brushless motor controller, which in turn is compatible with multiple brushless motors.
As is explained in the White Paper, using RDFox, knowledge graph experts can easily import new data along with a set of compatibility rules and the system will automatically reflect the changes in the data, thus always giving end users access to the latest data. Similarly, data can be deleted or the ontology can be adjusted to reflect changes in relationships, and the system and the user experience created with metaphactory will update accordingly.
Often there is also a need for end users to be able to modify or augment the data in the knowledge graph. Using metaphactory’s intuitive semantic forms, end users are able to seamlessly add new components to their catalogue:
Similarly, end users can change the configuration of existing components and the updates will immediately be propagated throughout the entire user experience. Let’s say, for example, that our support engineer from before is again looking for a rotation solution with specific parameters but within a predefined budget of €50. As depicted by the screenshot below, all of the rotation solutions in the system exceed the €50 budget.
After some negotiations with her supplier of DC motors, our engineer can adjust the price for the DC motor that should be included in the customer solution:
This will result in multiple rotation solutions being updated accordingly:
You can also find this blog on the metaphacts website.
Thank you to the team at metaphacts for making this Joint Solution possible.
The illustration below outlines the anatomy of a standard customer project based on metaphactory and RDFox. We guide and support customers through the entire process of building a FAIR data, knowledge graph-driven application that fits their needs. This covers core steps such as identifying main use cases, domain modelling and defining the application logic, knowledge graph creation, crafting the user experience and user interface, and deployment into production. All of this happens in a very agile manner and in close collaboration and synchronisation with your team of experts and selected end users, who have the opportunity to provide feedback and help refine the end solution at every step along the way. Typically, the implementation of a joint metaphactory-RDFox project takes approx. 5 weeks, depending on the use case scope and complexity — an unimaginable timeframe for traditional solutions!
The team behind Oxford Semantic Technologies started working on RDFox in 2011 at the Computer Science Department of the University of Oxford with the conviction that flexible and high-performance reasoning was a possibility for data-intensive applications without jeopardising the correctness of the results. RDFox is the first market-ready knowledge graph designed from the ground up with reasoning in mind. Oxford Semantic Technologies is a spin-out of the University of Oxford and is backed by leading investors including Samsung Venture Investment Corporation (SVIC), Oxford Sciences Enterprises (OSE) and Oxford University Innovation (OUI).