The Whiteboard

Chemical Manufacturer: Faceted Taxonomies 

Capturing interrelations of information for relevance can be difficult, even with NLP. More often companies will seek to work in taxonomy space in their journey toward richer implementations of knowledge graphs for automation adoption. Our consulting services leveraged this approach to provide a foundation stepping stone as the company sought to bring inherent knowledge graph capabilities into their business. 

This global manufacturer had a sluggish system in place to comb through internet publications and look for key terms that might mark articles of interest to its divisions for competitive intelligence as a spawning point for innovative ideas. However, processes remained heavily manual and cumbersome. They realized that strong text matching and analysis was a missing component and decided to turn to taxonomies to mitigate and improve the process. 

Semantic Arts quickly discovered that the key to success was faceted taxonomies. We  worked with SMEs to determine what areas contained specific controlled vocabularies and  specialized terminology. As a starting point, Semantic Arts created a series of taxonomies  for each area for improved automation. Areas included: Products, Industries, Customers,  Capabilities, Manufacturers, Materials and Processes 

The tight focus of each facet allowed for SMEs and division experts to create very specific lists of terms. By using preferred labels and alternate labels (synonyms) for each, SA  enabled what could be recognized and matched in a desired internet corpus. Initial  implementation of the facets showed a higher level of matching to recognized terms of  interest than an NLP algorithm achieved, created a higher confidence in the significance of  the match, and left out many common or “stop” terms that the original method still picked  up. A start of efficiency was realized. 

Semantic Arts developed a more extended road map with the manufacturer to first refine  and bulk up the taxonomy lists based on continued implementation and analysis. By implementing, the client’s intent will be to apply a simple semantic layer to relate and interconnect the taxonomy facets. This ontology model will allow even richer inferencing  and matching of results based on relationships between terms (i.e., an article about a  specific product will imply the involvement of certain manufacturers even if they are not  explicitly mentioned). 

In this case, a small step in a focused step into taxonomy re-classification is helping to open more understanding about the broader benefit while allowing for faster delivery of more pinpointed research answers. In addition, building pipelines of connected unstructured information is consistent with organizational goals of harmonizing data for greater  strategic value. Divisions in other parts of the enterprise have taken notice and there is 

expressed interest in leveraging the unique re-usability and interoperability semantic capabilities enable after this initial pilot.

Investment Bank Case Study: Operational Risk 

In this major investment bank managing all the flavors of operational risk has become very balkanized. There are separate systems for process management, risk identification,  controls, vendor risks, cyber risks, outsourced risks, fraud, internal incidents, external incidents, business continuity, disaster recovery inter-affiliate risk and many more. 

To address, we were able to create an elegant ontology that captured all these aspects of  risk. We then (one-by-one) were able to extract and conform their existing information into this shared model. 

We managed to catch the re-write of a control library in mid-stream and get them to persist the key information directly to a triple store. The mappings have been ported into production, and we built (in TARQL) the capability to create a unified view of information systems that feed risk evaluation metrics. Additionally, an interactive graphics capability has been built directly on the triplestore for visualization across the risk portfolio.

Investment Bank: Resolution Planning 

This is one of the “too big to fail” banks, who are required by regulators to implement  “resolution planning” or as it’s known on the street a “living will.” The first few generations of the resolution plan were long on long textual descriptions of the nature of the interactions between various legal entities within the bank. 

Our sponsor recognized that the key to making a resolution plan workable is to make it data driven rather than document driven. Document-driven resolution plans are out of date as soon as they are written and require humans to read and interpret. While the firm,  as with most large financial services firms, consists of thousands of legal entities, there are  “only” a few dozen that are significant from a resolution standpoint. However, this is made more complex because hundreds of departments (may and do) have service relationships with their peers in other countries and time zones. Often these arrangements are tacit rather than spelled out, and even those that are written fall far short of the regulators desire to see specific mechanisms for controlling the work and assuring it gets completed. 

We based this project on the concept of Inter-affiliate Service Level Agreements. We designed an ontology of Service Level Agreements and in the course of four months iterated it through eight versions as we learned more and more about the specifics of getting a new system designed and built. 

In addition to (and in parallel with) the ontology development we built an operating system,  using our model driven development environment. We populated a triple store with data  sourced from many of their existing systems (HR for personnel and departments, finance for legal entities and jurisdictions, IT for applications, hosting and data centers and the activity taxonomy from the project we had performed the previous year). On top of this we built user interfaces that allowed managers to document the agreements that were in place between themselves and other departments in other legal entities. 

We completed the project in time to demo to the regulators and it is now being used as the basis for their go forward Resolution Plan.

LexisNexis: Enterprise Ontology 

We worked with this leading provider of legal and medical knowledge to build an enterprise  ontology for their wide-ranging content. In addition to building an ontology for their case law and statutory product lines, we worked with their Master Data Management Initiatives.  

They have over 30 MDMs in various stages of development with logical data models. These models (and therefore the MDMs themselves) were integrated manually, in a somewhat ad-hoc fashion. We built tooling to convert their existing logical models into a single integrated ontology, where the integration points were far more obvious. From there, we built tooling to convert the ontology back to a set of similar, but now conformed, logical data models.

Management Consulting: Enterprise Search 

This major consulting firm has the enviable problem of having every possible desirable expertise characteristic’s somewhere within their ranks of 300,000 employees, and the unenviable problem of trying to find those needles in such a gigantic haystack. 

It’s not that they are unaware of the problem. They have launched many projects over the years to address this, some of which cost hundreds of millions of dollars (and a problem easily worth this much to solve, but very small increases in chargeability or win rates on proposals as a result are worth that on an annual basis). 

We built an ontology to integrate projects and proposals around expertise and proficiency.  We have harvested as much as is known about current employees in terms of skills and proficiency (and we are beginning to get subcontractors and partners), but we know that this information is not being kept up to date. We are at the early stages of two more initiatives, one that will nudge people to update their profiles when it becomes known that there is demand in a particular area; the other, to combine externally available information with this primarily internally sourced graph. 

The other side of this project is to replace the game of telephone that is currently the primary way to find key people in the firm. Currently, senior staff or partners rely on their network to find experts. Junior people are more often left out. In either case the process is quite haphazard as each request gets forwarded on to another subset of the network.  There is a great deal of reluctance to “spam” their internal network, but there is also the need to find the right people as rapidly as possible. 

We have built an early prototype model with the vision that a chat-based service will leverage the graph network as well as keep track of the results, such that over time the requests will get smarter, smaller, and resolve faster.

With a successful 200-year track record in developing extensive and diverse product lines by leveraging chemistry and science, this global innovator was embarking on a new era of discovery to shape a better world. 

Categorization of products and the relationships between those named entities were difficult to describe in traditional taxonomies and “systems of systems”. Over the course of  2 centuries massive amounts of complex information had been arrogated, however it didn’t give a full picture. The patterns of data intersection for decision-making were lost within the siloed systems. For greater predictable business insights based not only on structured data but unstructured documents the need for easier access and interoperability was a primary goal. 

A CoE (Center of Excellence) was formed to accelerate R&D development, analytics and  finding of data to advance innovation with greater speed and reliability. A secondary mission was to socialize this capability to the broader community. Ontology development and tools to harmonize information were a foundational part of this data enrichment strategy, but in-house skills and data-centric modeling expertise were insufficient. It was necessary to develop as a core competency. 

Semantic Arts consultants were engaged to bridge this critical ontology and semantic capabilities gap. Our strategic advisory service offering brought over 25 years of practical implementation learnings and educational workshops to deliver a series of focused topics that met broad expectations of a teaching library. Recorded videos are now available on  the company’s enterprise intranet to traverse the complexities of information silos by enabling knowledge graphs, ontologies, and data-centric thinking. 

Topics include – Introducing Semantic Technologies and Ontologies, Introduction to OWL,  Introduction and Hands-on with Protégé and Property Rules, Understanding Class Relationships,  Expressions, and Property Restrictions, Semantic Triple Visualizations, Ontology vs. Taxonomy,  Topic Extraction and Weak Signals, and UI Visualization with Knowledge Graphs 

In parallel Semantic Arts are engaged with a division focused more on enriching and unifying 19 different data repositories into an ontology. The goal is to model relationships between the concepts, substances, and components for describing the products in a disambiguous manner to eliminate data duplication, increase metadata clarity, and eventually incorporate ML and NLP capabilities. We’re collaborating with the client to structure a roadmap for implementation. A projected iterative, agile approach by using our  predictable (rinse and repeat), Think Big / Start Small methodology will be employed to  guide and instruct in this digital evolution.

This effort to interconnect information assets for discovery complements and aligns with the broader digital transformational for bringing “miracles of science” to realization.

Semantic Arts worked with PwC to improve Verizon’s data privacy capabilities. 

Prior to the engagement, as applications evolved many of the process steps required to stay in compliance with privacy regulations and policies were performed manually using  multiple data sources. Our goal was to provide a knowledge graph as a single source of privacy metadata (information about data classified as private). 

Early in the engagement, we identified the key components of the data privacy landscape,  summarized in this diagram: 

The data set loaded into the knowledge graph identifies which applications and third parties are involved in each kind of privacy data processing: data collection, analysis,  modification, transfer, storage, etc. 

We modeled most of the privacy concerns implied by the diagram, including: legal rights of  a data subject, agreements with third parties, details of data processing relevant to privacy  concerns, data lineage, data retention, data catalogs, impact assessments, privacy-related  processes and tasks 

We also created an extensive taxonomy that allowed key distinctions to be made while the core model remained simple and straight-forward. Finally, we created generic configurable queries to perform a range of data validations. 

The result is a consolidated view of data from multiple sources that allows greatly improved management of application compliance with privacy data policies and regulations.

International Monetary Fund

The IMF works to achieve sustainable growth for its approximately 200 member countries.  It carries out missions and loans funds to execute projects for the member countries. The countries’ financial situation is measured and tracked using a wide variety of economic indicators. 

The challenge: Information is stored in a wide variety of vocabulary managers, applications and databases. This makes it difficult to quickly get answers to questions required to carry out day-to-day work. For example: Who is likely to be an expert on customs; Find documents associated with countries similar to Afghanistan; What types of missions for what countries are addressing climate change? 

In each case, getting the answer requires retrieving and processing information from multiple sources. To represent the information required to answer the questions, we built an ontology that covers the core business of the IMF. The main things are: 

We created a knowledge graph composed of the ontology and RDF triples data that was created by converting taxonomies and datasets from a variety of data sources. We wrote  SPARQL queries that traverse the knowledge graph to answer the questions of interest.  This is to be the basis for an internal knowledge portal for integrating structured data (such  as GDP per country) with unstructured data (such as country-specific reports on commodity prices).

World Minerals: ERP Conversion Feasibility  

World Minerals is the largest producer of diatomaceous earth and a major producer of several other industrial materials. The infrastructure upon which their internally developed ERP system  was based had become obsolete. This included pretty much the whole stack: DEC VAX operating system, Alpha chips, the Rdb database, Cobol and green screens. 

They were contemplating implementing a packaged system when they contacted us. After a great  deal of research, we concluded: 

Because their current systems had some very complex built-in requirements that no ERP package supported, a package system would require a great deal of customization and extension. An automated conversion to modern architecture would be feasible, more cost-effective and less disruptive. 

They opted to go the conversion route. They contracted with an implementation firm who did a partially automated conversion, but who did come in close to our estimate, and they have been  using their converted systems since. 

Washington State: SOA Design 

The long-range plan at L&I called for organizing their future application initiatives around shared services and shared messages on a message bus. 

In this project we created detailed requirements specs for the dozen major shared services and created an inventory of the key messages they would need to form the backbone of their Service Oriented Architecture. 

One of the interesting early wins was with their Accounts Receivable system. They had just started an AR project, and we convinced them to think of AR as a service rather than an application. They had discovered that 23 of their 200 applications had implemented AR  functionality and this project was intended to rationalize this. The pilot application to be converted (Claims Overpayment) wanted to implement an AR message that was highly specific to claims overpayment and, therefore, would not be reusable. This was contrary to the idea of reusable messages. 

We did a bit of semantic modeling to find the commonalities and differences and constructed a common message that had variable payloads for the few fields that really needed to be specialized for each case. On a follow-up visit several years later, they  reported they have successfully converted all 23 of the satellite AR functions, which has  provided benefits including consistent revenue reporting and a single place to check to see  if someone owes the Agency money before they pay them from their Accounts Payable  systems.