FIBO in the Semandcs World

FIBO in the Seman<cs World Dennis E. Wisnosky Founder, Wizdom Systems, Inc. It is 2013 – Do you know where your money is? ME! •  Degrees in Physics, EE, Management Science -­‐ all part <me on a real campus while raising a family •  5 years private sector, 10 years Gov, 20 years private sector, 5 years Gov and now Wizdom Systems Inc. -­‐ reinvented. Google or YouTube or LinkedIn has more than even I know about me. •  Asked in 2006, by the the Sec of Defense to build system that could solve the DoD “where is my money problem”, I accepted. LiXle Did I Know! 10/2/13 www.wizdom.com 2 A Bed Time Story! Once upon a <me, there were no standard defini<ons of financial terms and the financial ins<tu<ons could interpret the meaning of the rules and regula<ons of the industry each in their own way. Thank you Daniel Pink! 10/2/13 www.wizdom.com 3 Everyday, new financial instruments and transac<on types were invented. One day, major companies in business for many decades began to collapse and lead the world into general economic depression. 10/2/13 www.wizdom.com 4 Because of that, regulators struggled migh<ly to understand the condi<on of the world’s economy and it became clear that the companies themselves did not know their true financial exposure. 10/1/13 They did not know the Provenance the TRUTH, about their data! www.wizdom.com 5 Because of that, an effort was launched by the industry to develop a Financial Industry Business Ontology (FIBO) -­‐ a common vocabulary based on interna<onal standards, that would enable companies to beXer communicate within and among themselves and would enable regulators to perform meaningful oversight as required by laws. 10/1/13 www.wizdom.com 6 Un<l finally, the dual purpose of reducing the cost of manufacturing data required by law became de minimis and Congress and regulators were confident of the provenance of answers to their ques<ons of the industry. And, They All Lived Happily Ever aher! 10/1/13 www.wizdom.com 7 My DoD Experience -­‐ A Metaphor for FIBO Implementa<on (BMA Federa<on Strategy version 2.4a) 2005 Present Future (BOE Execu<on Roadmap) (BMA Technical Transi<on Plan version 1) DoD Strategic Mgmt. Plan (SMP) DCMO/CIO Policies BEA 3.0 Performance Measures CIO – DIEA, Segment Archi. CV & Primi<ves Ini<al BOE Arch. Fed. Experience Version 2.4a MDR Biz. Intelligence Federa<on Implementa<on Plan BOE Vision 2009 Roadmap: Architecture Governance Socializa<on Services Infrastructure Vision & Strategy Planning & Roadmap Infrastructure Governance BOE Service Enablement Data Integra<on Business Intelligence Common Vocabulary (Ontologies) Domains HRM/ Med FM Logis<cs RPILM WSLM/ MSSM Seman<c Informa<on Execu<on Data Sharing and BI Enablement Past Enterprise Standards RDF OWL SPARQL other Rules/Workflow DCMO/CIO SOA Implement. Strategy BEA 9.x & Beyond Security 2012 – Federal Always MCIO/DISA aking It RCloud eal! (Business Infrastructure) NCES/CES DBSMC/IRBs DCMO/DCIO; EGB; BECCM DWiz DoD DCMO BMA CTO & CA 8 We Are on this Path Now! The 5-­‐Star Model Tim Berners-­‐Lee and others -­‐ 2010 + Only the Source has the TRUTH! 10/1/13 www.wizdom.com 9 How 5-­‐Star Data Works ★ make your stuff available on the Web (whatever format) under an open license1 example ... ★★ make it available as structured data (e.g., Excel instead of image scan of a table)2 example ... ★★★ use non-­‐proprietary formats (e.g., CSV instead of Excel)3 example ... ★★★★ use URIs to denote things, so that people can point at your stuff4 example ... ★★★★★ link your data to other data to provide context5 example ... hXp://5stardata.info 10/1/13 How Open Linked Data Works! www.wizdom.com 10 What Could be More Simple! Courtesy David McComb 10/1/13 Can be Expanded Without Limit! www.wizdom.com 11 Can be Expanded Without Limit Courtesy David McComb 10/1/13 What I did Think of! www.wizdom.com 12 The need for a Common Vocabulary Benefit shown by Example! 10/1/13 www.wizdom.com 13 Common Vocabulary in Ac<on California University of Pennsylvania
hasName
DBpedia
(Wikipedia)
Dataset
DoDAF Wizdom.
hasTitle
Dennis Wisnosky
graduatedFrom
university
hasName
hasName
hasName
University of Dayton
book
University of Pittsburgh
writtenBy
person
hasName
DoD HR
Dataset
person
bornIn
Dennis Wisnosky
Washington
yearOfBirth
19XX
locatedIn
Pennsylvania
Wikipedia Data: Who wrote “DoDAF Wizdom”?
DoD HR Data:
Where was Dennis Wisnosky born?
What is who
the true meaning f this? Linked Data: Where was the
person
wrote
“DoDAF
Wizdomoborn?
5/6/2010
14
Enormous Efficiency in Opera<ons BIG DATA? (near) Exponential
connections (n2 – n) I want what I want, I want what I need, I want to know it is right, I want it when I need it. (near) Linear connections (2n-1) For the data -­‐ SAP says each point to point interface costs U$300,000 to build and maintain. Seman<c Web Technology reduces the number of connec<ons by orders of magnitude! Force Directed Graph
10/1/13 www.wizdom.com And! 15 Enormous Efficiency in Development For the Sohware – HP says, “Industry figures & case studies show between 40% to 80% reduc<on in effort and <me”! OK, I Solved the DoD Problem, and then! 10/1/13 www.wizdom.com 16 An Bigger Personal Problem! My War was Over! Or was it! What had Happened? 10/1/13 www.wizdom.com 17 What had Happened? Many causes for the financial crisis have been suggested, with varying weight assigned by experts.[14] The U.S. Senate's Levin–Coburn Report asserted that the crisis was the result of "high risk, complex financial products; undisclosed conflicts of interest; the failure of regulators, the credit ra<ng agencies, and the market itself to rein in the excesses of Wall Street".[15] The 1999 repeal of the Glass-­‐Steagall Act effec<vely removed the separa<on between investment banks and depository banks in the United States.[16] Cri<cs argued that credit ra<ng agencies and investors failed to accurately price the risk involved with mortgage-­‐related financial products, and that governments did not adjust their regulatory prac<ces to address 21st-­‐century financial markets.[17] Research into the causes of the financial crisis has also focused on the role of interest rate spreads.[18] 10/1/13 www.wizdom.com 18 What lessons have we learned from Lehman? Traders on the floor of the New York Stock Exchange on Sept. 15, 2008, the day of the collapse of Lehman Brothers. (Andrew Harrer, Bloomberg / September 10, 2013) 10/1/13 hXp://www.chicagotribune.com/business/columnists/ct-­‐biz-­‐0915-­‐
phil-­‐20130915,0,5125535.column www.wizdom.com 19 The Feds Act! Dodd-­‐Frank nearly 3000 pages Must be Interpreted by: OFR SEC 10/1/13 CFTC OCC FRB ECB www.wizdom.com 20 The Industry Cries Foul! “The need to create useful data rather than just lots of data comes as large global ins<tu<ons face expenditures ranging from $150 million to $350 million each to comply with new post-­‐
credit crisis regulatory requirements in the United States, Europe and elsewhere. That is "significantly larger" than the level of expenditures required previously for complying with Sarbanes-­‐Oxley Act, Markets in Financial Instruments Direc<ve a
nd B
asel I
I r
equirements, Javier Perez-­‐Tasso, head of marke<ng at SWIFT. from before the crisis” The Industry Responds! 10/1/13 www.wizdom.com 21 What is Needed! • “A simplified and replicable method of calcula<ng exposure to risk that can be universally applied to sources of transac<ons that are reconcilable to accoun<ng records • Global iden<fica<on standards for legal en<<es, products and financial events to facilitate the aggrega<on and comparison of risk exposure data within and between financial ins<tu<ons and across the industry •  A ‘Big Data’ framework that is able to provide regulators and others with complete and accurate real-­‐<me informa<on rela<ng to the global financial system” This needs statement is summarized by the Basel CommiXee as “an intelligent seman<c network for systemic risk analysis.” Basel CommiXee on Banking Supervision in June 2013 issued for comment a paper en<tled Supervisory Framework for Measuring and Controlling Large Exposures. hXp://www.bis.org/publ/bcbs246.pdf 10/1/13 www.wizdom.com The Problem 22 Must Solve Problems Like This 10/1/13 David Newman Wells Fargo www.wizdom.com 23 What Do You See? Is it ok for different people to interpret this picture in different ways? 10/1/13 www.wizdom.com 24 What Should You See? Is it ok for different people to interpret this picture in different ways? Of course not! The meaning = seman<cs – of the data must be precisely defined. 10/1/13 Iden>fy Key Contractual Events Iden>fy Key Contractual Ac>ons Classify Contract Type by Cash Flow www.wizdom.com Default Events Increase Collateral Transfer Payments 25 Onto Physics “Physicists Debate Whether the World Is Made of Par>cles or Fields-­‐-­‐or Something Else En>rely” This ar>cle was originally published with the >tle What Is Real? Scien<fic American July 24, 2013 “It stands to reason that par<cle physics is about par<cles, and most people have a mental image of liXle billiard balls caroming around space. Yet the concept of “par<cle” falls apart on closer inspec<on.Many physicists think that par<cles are not things at all but excita<ons in a quantum field, the modern successor of classical fields such as the magne<c field. But fields, too, are paradoxical.If neither par<cles nor fields are fundamental, then what is? Some researchers think that the world, at root, does not consist of material things but of rela>ons or of proper>es, such as mass, charge and spin” 10/1/13 www.wizdom.com 26 – iden<fica<on of legal en<<es, their jurisdic<ons and ownership control hierarchies – Iden<fica<on of financial contracts and instruments – classifica<on and data linkage for aggrega<on – ac<onable risk intelligence Requires Par<cipa<on at all Levels – MORE 10/1/13 www.wizdom.com 27 Enterprise Data Management Council (EDMC) FIBO Par<cipants   Wells Fargo chairs the EDM Council’s Seman<c Technology Program, interfaces directly with regulatory authori<es and leads the working group that is responsible for construc<ng the opera<onal capabili<es of FIBO Deep Technical Dive! 10/1/13 www.wizdom.com 28 Conceptual Level – Abstrac<on is necessary to preserve Opera<onal Op<ons Level of abstrac<on Refinement and Specifica<on •  Opera<onal Level – Precision is necessary to ensure consistent and compliant execu<on Technical Dive! 10/1/13 www.wizdom.com 29 FIBO Business Conceptual and Opera<onal Ontologies are Two Sides of the Same Coin • 
FIBO Business Conceptual Ontologies   Primarily human facing   Visual blueprint   Standard terms and defini<ons for business concepts   Broad based expressions of conceptual specifica<ons, provenance, linkage and context of business constructs   FIBO Opera>onal Ontologies   Primarily machine facing (RDF/OWL)   Derived from FIBO Conceptual Ontologies   Op<mized for performance and scalability. Fewer abstrac<ons. Inferred rela<ons, mappings.   Classifica<on, data linkage, valida<on and seman<c query.   Deliver executable func<onality to fulfill use cases, enable data linkage, transparency and risk analy<cs 30 Intersec<ng Ontologies: Conceptual and Opera<onal   Use case specific classes, proper>es   Use Case neutral   Classes and proper>es   Meaning expressed in the “Language of the business”   Defini>ons   Namespaces   Op>mized for opera>onal func>ons (reasoning; queries)   Formally grounded in legal, accoun>ng etc. abstrac>ons   Annota>ons   Addi>on of rules   Mapping to other OWL ontologies 31 Why Conceptual Ontology? •  Why do we need ontology? –  Full and factual representa<on of business concepts –  Opens the way for reasoning based applica<ons and seman<c querying of data But… •  Why we need a conceptual ontology? –  Provides basis for an opera<onal ontology –  More detailed and nuanced view of the business –  Let me take you through a couple of scenarios… 32 Regulatory Repor<ng Current State ?
Reports (forms) FORMS REPORTING ENTITY FORMS REGULATORY AUTHORITY 33 Regulatory Repor<ng Current State ?
Reports (forms) FORMS FORMS REPORTING ENTITY Change in Repor<ng requirements = • 
Redevelopment effort REGULATORY AUTHORITY Uncertainty • 
Content of the reports • 
By each repor<ng en<ty • 
Are we comparing like with like? • 
For each system and form • 
Data quality and provenance 34 Regulatory Repor<ng with Seman<cs !
Thing Contract IR Swap REPORTING ENTITY Data is mapped from each system of record into a common ontology CDS Bond REGULATORY AUTHORITY Receives standardized, granular data aligned with standard ontology (FIBO) Reported as standardized, granular data Uses seman<c queries (SPARQL) to assemble informa<on Agnos<c to changes in forms Changes to forms need not require re-­‐
engineering by repor<ng en<<es 35 What the Bankers and Feds Will See Informa>onal Sharing across Regulatory Agencies Financial Ins>tu>ons Legacy Database(s) Trading Trading & System Compliance System(s) Mapping Mapping Seman>c Informa>on Integra>on Pla\orm Swap Trade & Regulatory Repor>ng Mapping Regulatory Agencies Legacy Database(s) Mapping Swap Data Repository Database(s) FpML Seman>c Network Graph Analysis Mapping Seman>c Informa>on Integra>on Pla\orm Regulatory Risk Analyst Legal En>ty Data Provider(s) Ins>tu>onal Risk Analyst Seman>c Triple Store Seman>c Triple Store Ontologies David Newman Wells Fargo 10/1/13 Ontologies Trusted Financial Linked Data Cloud Deep Technical Dive! www.wizdom.com 36 EDMC/FIBO Generic Use Case Conceptual Ontology Data Model Opera<onal Ontologies Data Models Built by operators using Web Protégé and other tools obeying the EDMC FIBO Founda<ons Standard Extends Instan<ates Type Data Ontology Instances Managed in a Hybrid Cloud by the Operators Represented As Data Ontology Schemas Represented As Extends FIBO Conceptual Ontologies IsA Uses Type Process Ontology Instances Type Instan<ates 8/28/13 Built by SME’s using UML and Web Protégé obeying the EDMC FIBO Founda<ons Standard Process Models BPMN-­‐2.0 Ontologies IsA Process Ontology Schemas Hosted on a EDMC managed Server Published by EDMC as a FIBO Standard Built According to EDMC/FIBO Standards Other Standards Dennis E. Wisnosky and Mohamed Keshk What the Operators Will See Conceptual Ontology Data Model Opera<onal Ontologies Data Models Extends Instan<ates Type Applicant Data Ontology Instances Applicant Data Ontology Schema Represented As Extends FIBO Conceptual Ontologies IsA Uses Type Loan Process Ontology Instances BPMN-­‐2.0 Ontologies Type Instan<ates 8/28/13 Represented As Make Disbursement Sub Process IsA Process Domain Ontology Schema IsA IsA Hosted on a EDMC managed Server Find All Model Instances for All Applicants Who Applied in 2011 and Were Disbursed By Check? IFW Ontology Schema What Really Happens! Dennis E. Wisnosky and Mohamed Keshk Find all process model instances for all applicants who applied for loans in 2011 and had disbursement by check Applicant Data Model Loan Instance Data LOAN DATA Make Disbursement Sub-­‐Process Baby Steps! FIBO Guarantees Sausage Making Business En<<es Founda<ons Status Today! 10/1/13 www.wizdom.com 40 FIBO Business En<<es OMG-­‐EDMC – Drah Spec FIBO Business En>>es Ontology En>ty Types Legal Persons Formal Organiza>on Corpora>ons Partnerships Trusts Ownership Control By Func>on Legal En>ty ID •  Types of corporate structure •  Organiza<onal hierarchies / rela<onships Michael BenneX EDMC Michael BenneX EDMC 41 FIBO Founda<ons OMG-­‐EDMC Drah Standard FIBO Founda>ons Ontology Common business modeling framework Common rela>ons Roles Goals, Objec>ves Agents, People Organiza>ons Agreements Law Control, Ownership Accoun>ng Loca>on Road Map! •  FIBO Founda<ons provides the basic conceptual “Glue” •  Common abstrac<ons grounded in law and business Michael BenneX EDMC 42 FIBO Provisional Roadmap
2013
Q2
2014
Q4
Q1
Industry
review
OMG finaliza<on
Q3
FIBO-­‐Founda>ons Global Terms and modeling framework Q2
Industry review
FIBO Business En>ty Domain ontology Q3
Q4
Final OMG finaliza<on
Final FIBO Securi>es Industry review
OMG finaliza<on
FIBO Deriva>ves Industry review
OMG finaliza<on
Domain ontology Domain ontology FIBO Loans Domain ontology Industry review
OMG finaliza<on
Final Final Final FIBO Market Data, CAE, Por\olio, Payments DD
omain FIBO MOther arket ata, oCntologies AE, Risk/Repor>ng DD
omain FIBO MOther arket ata, oCntologies AE, Risk/Repor>ng ntologies FIBO MOther arket DDomain ata, CoAE, Risk/Repor>ng Other Domain ontologies 43 Thank you! Ques<ons? [email protected] 10/1/13 www.wizdom.com 44 hXp://www.youtube.com/watch?v=OzW3Gc_yA9A Enterprise Architect -­‐ TOGAF / Banking Apply for Enterprise Architect -­‐ TOGAF / Banking Job Portal: IT / Infrastructure Func<on: Development Ref. No.: EA848D Loca<on: London -­‐ City Contract: Full-­‐Time PERMANENT posi<on Contract details: Bonus + Pension + Car Allowance Salary: Bonus + Pension + Car Allowance 10/2/13 www.wizdom.com 46