 So the work that I'm about to present to you is actually the work that I did as part of my master's education is taking the basics of FIBLE and its potential use cases. So I'll skip through the first two slides since David did a great job this morning giving the background. And again the presentation objective is to add value to the FIBLE team by doing independent evaluation using the techniques that they teach at Ontology Engineering Corsate, the University of Toronto, and also to share some of my insight as to how an industry may be able to use FIBLE and what the implementation models may look like. So the Ontology Evaluation. So at the University of Toronto there are two evaluation criteria that are taught in terms of how to evaluate Ontology. First one is the intrinsic evaluation which is evaluating the Ontology with respect to the axioms of the Ontology itself. So that's to test if the Ontology is self sustainable meaning that if it's consistent and complete and performs all the intended functionalities like within the Ontology model itself. The second type of evaluation is the extrinsic evaluation which is basically evaluating the Ontology with respect to the application of the Ontology to the outside world. So that includes how the Ontology can be used to express different concepts or like ideas and also whether the testing whether the Ontology successfully fulfills intended use cases. So in this presentation I'll be doing the intrinsic evaluation by going through the the quickly touching base on the development cycle, the structure, the property, and doing the consistency and completeness check using PradaJ and UPSPOL pitfall scanner and this is done using the FIBLE Foundations module specifically and then I'll be doing the intrinsic evaluation by going through some of the example expressivity that you can articulate using FIBLE and sharing some of the use cases. Again how I think FIBLE can be used in a real industry setting and then in the discussion section I'll go through the potential implementation models and what I personally would like to discuss more like for me to gain more understanding of the FIBLE. So the FIBLE development that's done by the Enterprise Data Management Council is the business conceptual ontology which is the framework itself. So the FIBLE that's being developed by the EDN council does not contain any specific data elements that are specific to organizations who may be using FIBLE. So in order to use FIBLE business conceptual ontology that's developed an organization has to extend it in the form of FIBLE operational ontology and this process typically involves tagging concepts and data elements that are specific to your business settings under the the framework of finance business conceptual ontology that's developed by EDN council. So this is the FIBLE development it's supposed to say build, test, deploy and maintain and just to draw some picture here the intrinsic evaluation it like relates to the the test portion and extrinsic evaluation relates to the the deployment for portion. So the FIBLE structure so the FIBLE is written in two formats RDF slash OWL and also UML XMI formats and FIBLE is created in multiple modules and I believe independent groups actually work on different modules and right now three modules are posted as standards on the OMG website namely FIBLE foundations, business entities and indices and indicators and I'll be using the FIBLE foundation modules to walk you through the consistency and completeness check using PradaJ and PFO scanner. So FIBLE has modular property so the table that you see here lists the sub modular groups and sub modular concepts that are captured as part of the FIBLE foundation module itself. So just the FIBLE foundation module is made up of these very elementary like data elements that are necessary to describe a higher level of FIBLE module concepts and data elements and in the next slide you will see that some of the sub module like dependencies so like the graph that you see on the top left corner the utility sub module that's the complete sub module describing the utility group which is made up of the three boxes which is the sub modular concepts and the the pecs that you see on the right side which does not which don't have boxes around them that's the another modular concepts that are imported into the utility sub module group in order to fully describe the utility sub module itself and on the bottom right corner you see the agents and people sub module like same idea but then the key difference that I wanted to show you is that some modular groups are simple like the top left corner where some some modular groups are quite complex as it requires to bring in like multiple other sub modular concepts in order to fully describe what it intends to describe. So in other words this is like a key property in my opinion because like the whole development like at the basis of it how FIBLE is developed is it's developed in a modular fashion so each module can be recycled so essentially you don't have to like try to recreate like the meaning or relationships when you're describing others all you have to do is just go back to the people person like organization and basically bring in the already existing relationships to describe more like higher level concepts. The slide that you see here is the consistent check that's done using Prodigy and Prodigy is a tool that many ontology engineers use to work on ontology files and it's convenient because when you bring in RDF ontology files into the platform it displays the data elements that are part of that particular document and if the file that's brought in is consistent that then you're like the whole data element list that you're gonna see will be displayed in black. If it's inconsistent that you're gonna see red data elements and as you can see everything is black and the file that I used to show you this is called FIBLE Foundations 1.0 which is kind of like the master file for the FIBLE Foundation module that actually imports all of the some modular groups that exist within the FIBLE Foundations module itself so it's kind of like the master file and I used the same file and put it in the the ontology pitfall scanner. So ontology pitfall scanner is based on the web so you can actually go to like a website and you can import the RDF ontology file and what it shows you is it shows you potential or not potential but like the pitfall errors that modeler may have made in creating the ontology file is what the website supports. So there are three types of pitfalls minor important and critical and the the main notion here is that if you have any critical errors displayed on the oops website that means your ontology is not gonna work whereas if it's minor or important like pitfall error messages that means the ontology itself will work it's just that like there are ways to like make it better so based on the results there were no critical messages so meaning that the FIBLE Foundations module that was developed and published as standard like is consistent and complete although there are some areas where like the tool suggests that like it could be made better. Now I'll move into on the extrinsic evaluation of FIBLE. So the FIBLE expressivity so the FIBLE can be used to describe like many different like business settings or questions so those include capturing subsidiary ownership relationships among corporations like where the company is located like who sits on the board of directors like what's the macroeconomic like conditions like you know like the like the rates unemployment rate and also like different types of players that exist in the finance industry like central bank different types of bank insurance companies and so on. So they're like literally like a lot of application areas where FIBLE can be used for as it is intended to be used. The first use case of FIBLE that came into my mind as I was like doing my project is because it's using like a common language you can use it to integrate heterogeneous data sources and construct a knowledge model that is specific to the user organization's business setting. So what that means is that like you can so there like God knows that like in the in the old relational database model like like different systems like working silos and they may like use different terms to describe like the same data elements. Using FIBLE you would be able to align them and like not only just align them like your organization whether it's within or outside of organizations will be able to look at the the model that's implemented and be able to like understand exactly like what each other have have done. So to check whether the model has been implemented properly I came up with two competency questions. So you can ask can outputs of a FIBLE operational ontology be translated by business units using the common FIBLE conceptual model and the second one is do you get the same end results using FIBLE driven systems integration model compared to using conventional relational database data processing method. The second use case of FIBLE could be as a decision support system. So because FIBLE is an ontology model like an ontology has like a unique strength in allowing users to get answers through ontology curing systems. So this is really like the main application area in my opinion how industry players may want to leverage FIBLE technology because using this decision support system they would be able to get answers to for example list all common equities by tier and list them by their position holdings in dollars. List all risk weighted assets by asset class show the capital requirement ratio change over the last three fiscal years and show the requirements to be classified as high quality liquidity assets. So like imagine that you have this like internal like Google you know like search engine like implemented in your own organizations that look at the data that's built using FIBLE as like the underlying technology and you can ask this type of questions and you can get answers. How cool would that be? Great. The last use case is using FIBLE as a semantic platform to extract and store knowledge about the industry setting. So this relates to the FIBLE that's the schema.org that they was talking about this morning. So once that schema like standardization has been implemented that means that you as an organization can leverage FIBLE to develop this semantic knowledge management system that scans the information coming from the web and actually collecting them in the way that the machine can understand what's stored within it. So for example like you could test if you can find the data elements associated with merger and acquisition articles that you saw on router.com or Bloomberg in your FIBLE driven semantic platform or you can ask the system to display all ownership changes of technology related companies that happened within the last 12 months. Again very cool concepts in my opinion. So for the discussion so I personally would like to study the process of converting FIBLE business conceptual ontology into operational ontology more because I still have some challenges trying to wrap my head around like what that process would look like or work like and I want to study if the FIBLE like when it's like completely developed is sufficient to like capture all sub-domain concepts and terms that exist within financial industry and then most importantly I think this is the key for bringing FIBLE into an organization like which is like completely new to it is to think about how FIBLE can be brought in and what the implementation model may look like because like all this organization like at least like majority of them still use the relational database systems and to go from relational database to like graph base which is what FIBLE is driven based out of is quite a big jump that you have to make so I thought of a couple or three suggestions how this may get completed. So the first idea is to replace the existing relational databases completely with triplets stores and operational ontologies so in other words complete remodeling the chances of this happening would be like practically zero because that means the bank would probably blow up. The second idea is to hardwire operational ontology with the existing relational databases so the user can manipulate the data stored in DMS through running direct queries through ontology. So this is using technologies like R2RML to basically like operate like your ontology model but the data that gets processed in the background is actually done using like the SQL codes or relational databases. Again I'm not sure like how like popular this would be because that may or may not compromise the existing systems and system capabilities. The last model which I think in my personal view is the most interesting model is bringing in new FIBLE driven semantic knowledge management system as a net new capacity added to your organization and also utilizing the existing relational databases world. So what that means is that as I mentioned in earlier slides there are very specific benefits that you can enjoy from a semantic driven knowledge management system and so you would bring in this new system that's driven by RDFL potentially by data or document database and and simply like put in all the data that's already structured in that fashion using five leveraging FIBLE to define the relationships and at the same time use the existing relational databases and like to do all the roll-ups as you already would and create like a summary results layer that gets converted into the no SQL format and get imported into this into FIBLE driven semantic knowledge management so you can fully utilize the benefits of implementing FIBLE driven system. So how to get there in my opinion you gotta be familiarized with the tools that support FIBLE implementation understand their various vendors and products out there that will allow you to do so. It's probably a good idea to get expert guidance on how to bring in the technology and also you can join FIBLE proof of concept or FIBLE implementation teams like to actually work with them to bring in the technology with the guidance into your organization. So with that I conclude my presentation do you have any questions. The question was for the consistency check like what consistency did I check? I believe that was a question yeah. So the consistency check that I that I tested for is when you look at this like knowledge framework like represented by FIBLE business conceptual ontology it has the it has the properties between different data elements and like in triple form so like subject pretty kid and object and the consistency check basically tested like if you say like A owns B and B owns C and if it says C owns A then essentially your like non-consistence is because A is supposed to own C so there is like some sort of conflict in terms of singularity yeah exactly yeah. Circularity. So those type of like if testing whether there is like a conflict in like like the properties or the the intended like the meaning that's put between like data elements like that's what's tested by the the consistency check yeah. So the question was the ontology if it gets returned critical error message on the oops pitfall scanner it's not gonna work so what do I mean by that? So to answer your question I I'm not sure what would work or what would happen actually like if you get the the critical error message because I was not able to run into that situation using FIBLE. Yeah so how I understand it is that if you get critical error message through oops pitfall scanner that means that like the pitfall scanner is saying the modeler made some foundational error or fundamental error that makes the again the ontology model inconsistent or there's some component missing.