 there was more. Yes, his topic today is industrial equipment reference data sets, a review of structures and utilities. So great. We look forward to hearing about it. We can see your presentation. Okay, good. Thank you. Thank you for inviting me to speak today. I will give a brief presentation about work that people who know me would normally not associate with me in engineering space. I'm involved in this because I'm one of the principal investigators on the ASC Center for Transforming Maintenance through Data Science. And I will be presenting today on behalf of my colleagues, Xiuzhao, Chau Ngyun, Dongmin, Belinda Hotkovich and Tyler Bikan. And I want to talk to you about industrial equipment reference data sets and our review of their structures and utilities. I don't think I need to tell you why semantic interoperability is something useful and what knowledge organization structures are. What I want to dive into a little bit today is how they are being used in equipment maintenance and what the lay of the land is and what we're wanting to use them for. So, after a brief introduction, I will look at them from the perspective of engineering standards, from the perspective of hierarchy equipment coding systems, and from the perspective of open source knowledge basis. So in equipment maintenance, there are lots of processes happening across organizations. There's a life cycle running horizontally through the organization. If you look at mining, then there's it starts in minerals exploration, there's the development of the mine, operation and maintenance and also marketing and sales. They all have their systems sitting next to each other in their silos of excellence, but they need to interoperate. And at the same time, there's also this vertical integration of the corporate perspective on to the operation into the divisions and eventually down to the pieces of equipment that need to be controlled and maintained. And if we want to move towards smart manufacturing, we need to have a much more intelligent way of linking between those. And just having this as human knowledge and human communications isn't enough, we have to have the machine support and machine support needs semantics. So, we looked at a long list of standards that exist that we found on the internet and literature research and this table gives you a little bit of an overview that there's a variety of structures, some just lists, you have taxonomies, terminology of lexical databases, so more and more organization theory and ontologies. And later on I will show you a bit more about how those are presented, but this already tells us that there are different levels of organizational structural maturity as you might call it. If you want to have more than just a list of terms. So in engineering standards, what we define them as existing standard models that relate to process plans for fracturing maintenance and electrical transmissions, because that's what we were interested in, in mining operations. And there's a couple of standards that relate to process plans and electrical power networks. There's some things around smart buildings, which the market themselves as building smart. Then there's very useful dictionaries from manufacturing, especially from automotive and aerospace they're very advanced in that. And then industry reference data libraries and dictionaries also from ISO. And this is a very, very busy table that lists those standards and what we then looked at we looked at who developed them for what purpose, how are they codified. And what do they describe, how many classes are there, and it's anywhere between two and 11,000. And then, and how many levels are they organized and when were they last revised because the last revision also tells us something about whether there's any active community involvement in developing them further. Then we have hierarchy equipment coding systems where hierarchy already implies that there's a bit more of a structure more than lists. And they are standards of classifying equipment with coding in hierarchical systems that are organized from from a top level perspective with a specific purpose in mind, but potentially quite useful for what we have in mind for organizing maintenance, because they're nicely structured. We looked at three of them the harmonized system of the world customers organization, which certainly is very much in use. And the common procurement vocabulary of the European Union, which is also very much in use and more detailed than the world customers organization and even more detailed very much what detailed is the electropedia of the International Technical Commission, which with 22,000 concepts and they are then also as because they're hierarchical organized in several levels. The interesting thing though is that the application space is very broad so the harmonized system of the world customers organization even though it has around 5000 codes is everything anything that goes through customs. So in terms of what we might find useful in maintenance. It's only a small fraction that actually is useful. Same in the procurement vocabulary with its 9000 terms, most of them are not applicable to to equipment maintenance. But the electropedia then has plenty and you can describe everything in quite fine detail. And then we have the open source knowledge basis where what we were aiming for were publicly available knowledge graphs. And there's concept net and wiki data that we looked at in more detail. Concept net is a knowledge graph, which occasionally gets released and you update with 1.5 million entities encoded in it. It's mainly used by natural language processing and linguistic researchers, computer vision researchers and engineering researchers and then there's wiki data which is extracted from Wikipedia. It's used by researchers and engineers with more than 100 million entries, so the very rich knowledge basis that we can use for for our work. So what does this work about in the Center for Transforming Maintenance through data science, we have a project on knowledge organization with the cute name koala. And this project is about information extraction and reasoning to support industry decision makers. What we do here is that we use natural language processing or technical language processing as a sub discipline of that to analyze work orders and maintenance reports. We find out what the status of a piece of equipment in the plant is, whether there's need for maintenance or risk of failure. And so in this project we've built a couple of engineering specific tools and pipelines for information extraction to display that information and be able to query that. And this is where the knowledge crafts come in because they allow us to traverse across all those documents and find the relationships of things, how they relate to say failure modes, how a piece of equipment is connected to another one. And it even allows us to do quite sophisticated reasoning towards understanding failure modes where the failure of one piece of equipment can result in a failure of another piece further downstream, which then can result in a catastrophic cascade of failures and bring everything to a halt. We didn't want to reinvent the wheel, obviously so koala identifies existing equipment taxonomies to augment and organize queries on our knowledge class that we built from all this material. So this was a very quick overview of the koala project I don't understand between you and the tea break and thank you for your attention.