Ontolog Forum

Ontology Summit 2016 Semantic Integration in Finance - Thu 2016-03-24

Session Co-Chairs: Mike Bennett and Andrea Westerinen


  • Overview - Brief overview of Semantic Integration Issues in Finance by Mike Bennett
  • Speakers:
    • Elisa Kendall (Thematix Partners LLC) Using Business Architecture and Semantics to Drive Data Quality Improvement in Banking
      • Elisa Kendall of Thematix will present on the use of formal ontologies in data quality improvement at a systemically important European bank.
      • slides to follow
    • Michael Uschold (Semantic Arts); Lynn Calahan (Wells Fargo) Referencing the Home Mortgage Disclosure Act in Building a Loans Ontology
      • This presentation describes the effort to represent in the Financial Industry Business Ontology (FIBO) Loans ontology the newly revised Home Mortgage Disclosure Act (HMDA) rule recently promulgated by the US Consumer Financial Protection Bureau. We will review how we use HMDA as a key use case to drive out requirements for FIBO Loans for the purpose of interoperability among loan processing applications, systems and databases. We explain our approach, starting from consuming and interpreting the Rule, relation to other standards, and through the creation of ontology concepts in OWL.
      • Slides
    • MikeBennett (Hypercube); Juan Sequeda (Capsenta) Creating a Virtual Knowledge Base for Financial Risk and Reporting
      • This presentation will demonstrate a practical architecture by which existing bank data can be queried using semantic queries and semantic-to-relational adapters, effectively creating a virtualized ontology-based knowledge base that exposes a range of existing data sources to semantic querying.
      • Slides Slides in pptx format
    • David Saul (State Street Corporation) Do you know where your data is? How FIBO Makes Data Smarter and More Governable
      • This presentation will focus on how the Financial Industry Business Ontology (FIBO) has been used at scale to integrate data from multiple sources, including instrument and business entity reference data.
      • Slides

Conference Call Details

  • Date: Thursday, 24-Mar-2016
  • Start Time: 9:30am PDT / 12:30pm EDT / 5:30pm CEST / 4:30pm BST / 1630 UTC
  • Expected Call Duration: ~2 hours
  • Dial-in:
    • Phone (US): +1 (425) 440-5100 ... (long distance cost may apply)
      • ... [ backup nbr: (315) 401-3279 ]
      • when prompted enter Conference ID: 843758#
    • Skype: join.conference (i.e. make a skype call to the contact with skypeID="join.conference") ... (generally free-of-charge, when connecting from your computer ... ref.)
      • when prompted enter Conference ID: 843758#
      • Unfamiliar with how to do this on Skype? ...
        • Add the contact "join.conference" to your skype contact list first. To participate in the teleconference, make a skype call to "join.conference", then open the dial pad (see platform-specific instructions below) and enter the Conference ID: 843758# when prompted.
      • Can't find Skype Dial pad? ...
        • for Windows Skype users: Can't find Skype Dial pad? ... it's under the "Call" dropdown menu as "Show Dial pad"
        • for Linux Skype users: please note that the dial-pad is only available on v4.1 (or later; or on the earlier Skype versions 2.x,) if the dialpad button is not shown in the call window you need to press the "d" hotkey to enable it. ... (ref.)
  • In-session chat-room url:
    • instructions: once you got access to the page, click on the "settings" button, and identify yourself (by modifying the Name field from "anonymous" to your real name, like "JaneDoe").
    • You can indicate that you want to ask a question verbally by clicking on the "hand" button, and wait for the moderator to call on you; or, type and send your question into the chat window at the bottom of the screen.
    • thanks to the folks, one can now use a jabber/xmpp client (e.g. gtalk) to join this chatroom. Just add the room as a buddy - (in our case here) ... Handy for mobile devices!
  • Discussions and Q & A:
    • Nominally, when a presentation is in progress, the moderator will mute everyone, except for the speaker.
    • To un-mute, press "*7" ... To mute, press "*6" (please mute your phone, especially if you are in a noisy surrounding, or if you are introducing noise, echoes, etc. into the conference line.)
    • we will usually save all questions and discussions till after all presentations are through. You are encouraged to jot down questions onto the chat-area in the mean time (that way, they get documented; and you might even get some answers in the interim, through the chat.)
    • During the Q&A / discussion segment (when everyone is muted), If you want to speak or have questions or remarks to make, please raise your hand (virtually) by clicking on the "hand button" (lower right) on the chat session page. You may speak when acknowledged by the session moderator (again, press "*7" on your phone to un-mute). Test your voice and introduce yourself first before proceeding with your remarks, please. (Please remember to click on the "hand button" again (to lower your hand) and press "*6" on your phone to mute yourself after you are done speaking.)
  • RSVP to with your affiliation appreciated, ... or simply just by adding yourself to the "Expected Attendee" list below (if you are a member of the community already.)
  • This session, like all other Ontolog events, is open to the public. Information relating to this session is shared on this wiki page.
  • Please note that this session may be recorded, and if so, the audio archive is expected to be made available as open content, along with the proceedings of the call to our community membership and the public at-large under our prevailing open IPR policy.



[12:38] MikeBennett: Speakers: please use *7 to unmute yourselves when you are ready to speak.

[12:44] Mark Underwood: FYI is launched - The Events internal URL is

[12:45] John Sowa: I just looked through the animations. I can't speak for everyone, but I much prefer to see the entire picture at once instead of seeing a little piece at a time.

[12:46] John Sowa: I'm happier with the PDF than the PPTX.

[12:50] Mark Underwood: As always, Twitter hashtag #ontologysummit

[12:56] RaviSharma: Mike - what is the mechanism to relate conceptual to logical data models in this ontology context, same as RDBMS or different?

[12:59] Donna Fritzsche: who is doing the work on swaps/derivatives/etc? is this one group or a community?

[13:01] Donna Fritzsche: risk aggregation is a great use case for semantic interoperability

[13:03] gary berg-cross: We have submitted a proposal to include small set of general financial terms into the core vocabulary . More details of the proposal can be found on the W3C FIBO Community Wiki, and in a test version of that includes the proposed enhancements. We re asking for support by preparing a short post to the W3C FIBO Community mailing list (you do not need to be a W3C member). Thanks for your cooperation

[13:08] MikeBennett: @RaviSharma good question. UML trace links will work within a UML-based environment; for RDF I think SKOS is the way to go, but I recommend extending the native SKOS with specific kinds of broader-match and narrower-match to specifically reflect the logical to conceptual relationship (not an equivalence)

[13:09] MikeBennett: @DonaFritzche The swaps proof of concept is what David Saul is presenting now.

[13:10] John Sowa: Donna, I strongly agree that risk analysis is essential. But this gets into a huge number of semantic issues. Just defining 'risk' in general is nontrivial. And for each case, the source of risk is extremely complex and context dependent.

[13:11] John Sowa: Note that most of the highly knowledgable humans in 2007 were unaware of the complexity of the problem.

[13:11] MikeBennett: @JohnSowa quite so. There is not only data about risk but also data that the risk is about i.e. instruments, positions, exposures etc.

[13:11] John Sowa: Even today, nobody has a clue about how to define an ontology of risk.

[13:13] ToddSchneider: John, the NCOIC SCOPE model could be used to start at notions of 'risk' (i.e., it's many related dimensions).

[13:13] Mark Underwood: @MikeBennett I'm guessing these slides are not Creative Commons Sharealike - maybe should not host unless footers are changed?

[13:13] MikeBennett: The ontological definition of risk is easy: Risk = Probability x Impact. But impact on what (goals); what event e.g. default. Some people's risk event is someone else's opportunity event. So there is a lot of detail in classifying risks according to variations in each of these dimensions.

[13:13] Donna Fritzsche: Hi John, I worked at an options trading firm (later SwissBank) in the 90s - so I am well-aware of the complexities. (thus my statement!)

[13:14] MikeBennett: @MarkUnderwood I'll ask - it should be the case that they are shareable.

[13:15] Donna Fritzsche: The trading firm I worked at was one of the first mainstream users of Lisp Machines - and very sophisticated in their handling of options - so I had somewhat of an inside-view.

[13:15] RaviSharma: John - I agree, EU effort (Prof. Mark von Rosing) is in beginning efforts - working on that subject and have a concept draft paper that I can request to share. But it tries to address more than finance space and is introductory compendium of list of risk ontologies contemplated or encompassed.

[13:15] Michael Uschold: I disagree that "The ontological definition of risk is easy". Risk is a highly ambigous term. Mike mentions one of any number of valid and useful perspectives about risk. We worked with a majore investment bank and they could not reach agreement, so we banned the use of the term 'risk' in the ontology since it was guaranteed to be ambiguous, underming the purpose of an ontology.

[13:16] John Sowa: Todd, I agree that many people have worked on issues about risk for a long time. But there is a huge gap between a definition that human experts use and computer systems can process.

[13:16] MikeBennett: @MichaelUschold Risk is a highly ambiguous word. That's why I don't like words. There are concepts which are tractable, though none of them are simple.

[13:18] ToddSchneider: John, agreed (for many other notions also). But that's what the SCOPE model and application were intended to help mitigate.

[13:18] Mark Underwood: @Gary - this is the list, correct?

[13:18] Dennis Wisnosky: We have submitted a proposal to include small set of general financial terms into the core vocabulary .

More details of the proposal can be found on the W3C FIBO Community Wiki, and in a test version of that includes the proposed enhancements.

We re asking for support by preparing a short post to the W3C FIBO Community mailing list (you do not need to be a W3C member).

Thanks for your cooperation

[13:19] MikeBennett: @Mark the post by Gary above is actually from Dennis. See immediately above at hh:18

[13:19] Mark Underwood: @Mike OK thx

[13:20] Mark Underwood: @Dennis - that page is the proposal; the listserv is at, correct?

[13:23] RaviSharma: Mike - Saul- congratulations to all involved as the proof of concept results appear encouraging.

[13:24] Michael Uschold: Question to speaker: how did you produce the triples? Write scripts? From ANZO directly? R2RML?

[13:25] John Sowa: David Saul's slide 17 is a good summary of the state of the art. Current technology provides good classifications, pie charts, and diagrams. They might help a human expert navigate through the data, but a highly knowledgeable human must interpret them. They cab't by themselves determine risk.

[13:25] Robert@MakoLab: Here: there is FIBO core proposal for based on FIBO with support of the W3C FIBO Community Group (

[13:27] Dennis Wisnosky: Pl go here to leave a positive comment about

[13:28] Michael Uschold: ON risk, see:

[13:28] Mark Underwood: Ack'd - Thx for the sharing go-ahead

[13:31] Mark Underwood: Headbanging collision of roles: big data (unstructured data lake-heads), MDM (tag-everything-heads) and BI (dashboard-heads)

[13:31] Donna Fritzsche: In terms of semantic interoperability - there is a problem with the idea that only "certified" ontologist (highly qualified) can understand the ontologies. This is a problem for the field in general.

[13:33] ToddSchneider: Donna, not really. As with any 'representative model' there will be complexity that not everyone wants to understand.

[13:33] MikeBennett: @Donna +1 - this is something we need to address with business-facing views (diagrammatic, tabular and natural language). For the original subject matter expert knowledge gathering we were able to explain the ontology content in basic set theoretic terms which anyone can understand. This did not cover the more complex logic in restrictions (though much of that was explicit in how we conducted the reviews, e.g. necessary conditions).

[13:33] RaviSharma: Lynn - slide 4 does not mention mortgage insurance? is that being too granular?

[13:35] John Sowa: Mike, I agree that it's important for people to agree on definitions. But that is just the earliest starting point for analysis. It doesn't "connect the dots" in the data, it doesn't do the inferences, it doesn't provide "actionable intelligence", and it doesn't suggest what actions to take about that intelligence.

[13:36] Donna Fritzsche: Yes Todd - but one should not have to be an Dean or Eliza to contribute, understand and further models - that are meant for widespread consumption. Agree with Mike that business-facing views are imperative. The rules around derivatives are highly complex and need to be communicated clearly to an audience broader then a few ontology experts.

[13:36] ToddSchneider: As an engineered artifact, an ontology is subject to the same pitfalls as other such artifacts: Providing sufficient 'metadata' that provides explanation, assumptions, provenance, etc. in natural language.

[13:36] Donna Fritzsche: +1 John's points

[13:37] MikeBennett: @John agreed - a first step is getting consensus on concepts, including the properties that connect them. If it's part of the real world it must be possible to present a view of the real world to the real people who work with it.

[13:37] MikeBennett: Elisa Kendall sends her apologies.

[13:40] ToddSchneider: Donna, the issue of addressing stakeholder concerns and there acceptance of proposed solutions is a common problem, nothing special to ontologies or their use.

[13:41] Donna Fritzsche: @Todd - agreed the communication tools (beyond metadata) are necessary. Natural language, visual representations, and annotated visual representations are useful.

[13:41] Donna Fritzsche: Todd - I would disagree in that we are discussing Semantic Interoperability - so communication above and beyond the norm is important.

[13:44] RaviSharma: Lynn - Does HMDA quarterly reporting imply removal of non- or- less performing loans from aggregations and collections to avoid 2008 uncertainties?

[13:45] ToddSchneider: Donna, yes. Interoperability is always a challenge (see the SCOPE model). Semantic interoperability is even more so. Part of the difficulty is that there can be a lack of awareness of the scope of ones semantics (e.g., vocabulary and interpretations) in comparison to the boundaries that will need to be crossed as part of interoperation.

[13:45] MikeBennett: My apologies to Lynn, I should have let them know we did not have screen sharing available.

[13:48] John Sowa: Lynn C's slide 6 summarizes the issues. How can data from all those "lines of business" be harmonized? No two banks that merge are able to merge their databases. They just keep running all the software from both banks indefinitely. Even in the same company, it's a major challenge to merge data from sales, engineering, manufacturing, research, and upper management.

[13:49] Donna Fritzsche: Todd - good points, re: elevating scope of semantics in comparison to the boundaries that will be crossed.

[13:54] RaviSharma: John - Yes this is a challenge, are data warehouses marts and big data stores able to ease this merge problem?

[13:55] Lynn Calahan: Ravi, since the HMDA Reg is based on applications, there is actually no information provided to the Regulator about performance for these loans. However, the Reg does include new identification requirements -- the LEI, a new Universal Loan Identifier--to assist in integration across the mortgage or loan lifecycle.

[13:56] BobbinTeegarden: @Mike U: how would you use/integrate SHACL?

[13:57] Michael Uschold: TBD, very eary thoughts on this.

[13:57] Lynn Calahan: John Sowa, that can be true. That is the type of use case--integration inside the bank--that we are trying to support. Also, for mortgage especially, the integration of documents and data for a process that is still especially paper driven by regulator and investor mandate.

[13:59] BobbinTeegarden: @MikeU Great idea, and probably other uses I would bet. Shapes of views of the ontology (depending on purpose or role...?)

[14:02] Michael Uschold: @JohnSowa said: "No two banks that merge are able to merge their databases." that seems to be empiracly true. We have a counterexample in the making, where we build an ontology, converted data to triples agains that ontology, then did the same exercise with an independently created DB form a company that had been acquired 10 years prior. IN ten years, they either never tried, or tried and failed to integrate the data.

[14:03] Michael Uschold: To @JohnSowa - to clarify, our example is not a bank, but the principle should hold in any industry.

[14:05] Lynn Calahan: Just a counterpoint: when banks merge, they sometimes do integrate systems. There are some distinct examples of this.

[14:06] ToddSchneider: Donna, (my interpretation is that) a goal of using ontologies in information system is to constraint interpretation. When creating or reusing an ontology attempting to extract a stakeholders (expected) interpretations can be problematic due to the vagueness with which many people use (natural language) terms. When there are stakeholders from different domains (i.e., crossing boundaries) reconciling the differences of vagarities explodes the problem.

[14:06] Lynn Calahan: I would agree with that statement, Todd.

[14:07] Donna Fritzsche: @John - Banks/trading companies that have merged do sometimes merge very large databases including financial options/multiple currencies/, etc. I have been part of this type of project. If it had failed (happened over the weekend - with several test merges happening in the month before) - it literally would have moved the markets.This type of project had a 3 year lifespan and involved semantic/syntactic merging. The person who ran the project is someone I greatly admire.

[14:08] RaviSharma: Lynn and Mike - Great I will try to reach you and thanks for great coverage of FIBO and Mortgage aspect.

[14:08] MikeBennett:

[14:08] Lynn Calahan: For people interested in participating in the Loans content team, I can be contacted at :-)

[14:08] Mark Underwood:

[14:09] John Sowa: Mike U, I just used banks as an example of a fairly conventional "line of business". If you go to "Web business", all bets are off. How could you relate Facebook, Twitter, Google, Amazon, etc., etc. Every effort to standardize anything leads to more standards -- e.g., compared to the Amazon database schema.

[14:11] Donna Fritzsche: @todd - you might have misinterpreted my poorly typed point. I did not propose that we elevate that scope (although my message might imply that.) I am still adjusting to this type of chat communication - my apologies!

[14:12] RaviSharma: Mike - I presume Basel 14 principles include - allows reserves based on Economic Capital Planning models?

[14:20] Jennifer Bond-Caswell: Have to drop off

[14:24] MikeBennett: @Ravi the BCBS239 principle are more about the reporting. Existing requirements and reports e.g. around reserves, should then be able to be reported under the new principles for data quality and timeleness and so on

[14:27] RaviSharma: Juan - Great does NoETL provide very low latency (or out of sync) with databases and takes care of access to last updated record? Another Q is how is R represented from data models (ER) when we do R2RML.

[14:28] RaviSharma: Juan - do you require time stamped data in multiple RDBMS SPARQL when you generate SQL?

[14:29] Michael Uschold: Question to @MikeB & @JuanC Is this talk about an actual proof of concept like David Saul described, where you have an actual company with actual financial data? Or, is it limited to being a clear description of how that can be done, in general?

[14:30] MikeBennett: This is what can be done. Keen to implement at a bank, it's a great fit to the BCBS239 requirement

[14:32] RaviSharma: Juan and Mike - it seems to me more pertinent to query determination of changes through metadata parameters (say from warehouse or a separate server) and then drill down and worry about synchronizing (temporally at least) different databases.

[14:33] ToddSchneider: Juan, the underlying 'value' is making 'information', data and/or metadata, explicit. Once explicit, it can be used or modified more easily.

[14:35] Mark Underwood: Have to run - Thx to speakers, and Mike - definitely interop happening with some fidelity

[14:35] MikeBennett: Thanks Mark.

[14:36] Dennis Wisnosky:

[14:37] Donna Fritzsche: Thanks Everyone!

[14:37] MikeBennett: Thanks to all our speakers and participants!

[14:37] RaviSharma: great session

[14:38] RaviSharma: thanks

Audio Recording