Enterprise Search: Pool Party and Philosophy 101

September 8, 2016

I noted this catchphrase: “An enterprise without a semantic layer is like a country without a map.” I immediately thought of this statement made by Polish-American scientist and philosopher Alfred Korzybski:

The map is not the territory.

When I think about enterprise search, I am thrilled to have an opportunity to do the type of thinking demanded in my college class in philosophy and logic. Great fun. I am confident that any procurement team will be invigorated by an animated discussion about representations of reality.

I did a bit of digging and located “Introducing a Graph-based Semantic Layer in Enterprises” as the source of the “country without a map” statement.

What is interesting about the article is that the payload appears at the end of the write up. The magic of information representation as a way to make enterprise search finally work is technology from a company called Pool Party.

Pool Party describes itself this way:

Pool Party is a semantic technology platform developed, owned and licensed by the Semantic Web Company. The company is also involved in international R&D projects, which continuously impact the product development. The EU-based company has been a pioneer in the Semantic Web for over a decade.

From my reading of the article and the company’s marketing collateral it strikes me that this is a 12 year old semantic software and consulting company.

The idea is that there is a pool of structured and unstructured information. The company performs content processing and offers such features as:

  • Taxonomy editor and maintenance
  • A controlled vocabulary management component
  • An audit trail to see who changed what and when
  • Link analysis
  • User role management
  • Workflows.

The write up with the catchphrase provides an informational foundation for the company’s semantic approach to enterprise search and retrieval; for example, the company’s four layered architecture:

image

The base is the content layer. There is a metadata layer which in Harrod’s Creek is called “indexing”. There is the “semantic layer”. At the top is the interface layer. The “semantic” layer seems to be the secret sauce in the recipe for information access. The phrase used to describe the value added content processing is “semantic knowledge graphs.” These, according to the article:

let you find out unknown linkages or even non-obvious patterns to give you new insights into your data.

The system performs entity extraction, supports custom ontologies (a concept designed to make subject matter experts quiver), text analysis, and “graph search.”

Graph search is, according to the company’s Web site:

Semantic search at the highest level: Pool Party Graph Search Server combines the power of graph databases and SPARQL engines with features of ‘traditional’ search engines. Document search and visual  analytics: Benefit from additional  insights through interactive visualizations of reports and search results derived from your data lake by executing sophisticated SPARQL queries.

To make this more clear, the company offers a number of videos via YouTube.

The idea reminded us of the approach taken in BAE NetReveal and Palantir Gotham products.

Pool Party emphasizes, as does Palantir, that humans play an important role in the system. Instead of “augmented intelligence,” the article describes the approach methods which “combine machine learning and human intelligence.”

The company’s annual growth rate is more than 20 percent. The firm has customers in more than 20 countries. Customers include Pearson, Credit Suisse, the European Commission, Springer Nature, Wolters Kluwer, and the World Bank and “many other customers.” The firm’s projected “Euro R&D project volume” is 17 million (although I am not sure what this 17,000,000 number means. The company’s partners include Accenture, Complexible, Digirati, and EPAM, among others.

I noted that the company uses the catchphrase: “Semantic Web Company” and the catchphrase “Linking data to knowledge.”

The catchphrase, I assume, make it easier for some to understand the firm’s graph based semantic approach. I am still mired in figuring out that the map is not the territory.

Stephen E Arnold, September 8, 2016

Comments

Comments are closed.

  • Archives

  • Recent Posts

  • Meta