PoolParty - Technical White Paper

Introduction

PoolParty® Semantic Suite (https://www.poolparty.biz/) is an AI platform based on semantic technologies and machine learning. It helps organizations to build and manage knowledge graphs as a basis for various AI applications. As a semantic middleware, PoolParty extracts the semantic meaning from your data and links your business objects and content assets automatically. Make your data actionable and benefit from smart applications!

PoolParty Technical Overview

PoolParty technology platform consists of several components and can be configured and extended to meet individual requirements. The components are grouped as outlined in the diagram below into four main areas:

  • Knowledge Graph Management
  • Text Mining & NLP
  • Linked Data Orchestration
  • Semantic AI Applications


PoolParty Semantic Suite - Components and Features

Knowledge Graph Management

PoolParty Thesaurus Management supports web-based taxonomy management. It is completely built on top of W3C’s Semantic Web standards (http://www.w3.org/standards/semanticweb/). In its core, PoolParty uses the Resource Description Framework (RDF) to represent SKOS vocabularies (e.g. taxonomies or thesauri). PoolParty is ready to consume and to publish Linked Data out-of-the-box. PoolParty offers web-based APIs and a SPARQL endpoint ( http://www.w3.org/TR/rdf-sparql-query/) to execute queries over each thesaurus project. This component can also be used to integrate knowledge graphs with content platforms (Wikis, CMS, etc.) or search engines.

PoolParty Ontology Management supports web-based ontology management. It is completely built on top of W3C’s Semantic Web standards and allows to model ontologies based on OWL 2 or reuse existing ontologies like FOAF, FIBO, Schema.org, etc. In addition, PoolParty introduces the principle of 'Custom Schemes' that allow to create views on subsets of a single ontology or to combine elements from different ontologies. Custom schemes can be used to extend thesauri, provide facets in GraphSearch, map relational data to RDF in UnifiedViews or as a basis to specify an editor in the PoolParty GraphEditor.

PoolParty GraphEditor supports web-based linked data management. GraphEditor allows to create custom views on RDF data based on custom schemes defined in the PoolParty Ontology Management. In addition taxonomies or thesauri defined in the PoolParty Thesaurus Management can be used as controlled vocabularies/metadata defining data in the GraphEditor. This component also supports bulk editing and lets data engineers interact with RDF data without the need of having deep knowledge in SPARQL.

Text Mining & NLP

PoolParty Extractor supports highly scalable and precise entity extraction, based on knowledge graphs as well as machine learning, which can be combined, put in series, or even used as parts of more complex rules and constraints for sophisticated text mining tasks. Its ability to transform structured and unstructured information into RDF offers new options for data analytics.

PoolParty Semantic Classifier works well together with the PoolParty Extractor while classifying whole text fragments or documents. It is based on machine learning algorithms like SVM, Deep Learning, Naive Bayes, and others. It is well proven that PoolParty Semantic Classifier is able to outperform other tools of this kind when using controlled vocabularies to label training documents based on an established domain knowledge model.

Linked Data Orchestration

PoolParty UnifiedViews supports automation of various data management tasks along the whole Linked Data Life Cycle. Typical tasks fulfilled by UnifiedViews are data ingestion, data transformation, enrichment, entity linking, or data quality assurance. UnifiedViews provides a large library of Data Processing Units (DPUs) that can be used as parts of whole data processing pipelines. Configuration of such pipelines can be managed in a user-friendly way by using a graphical editor. Pipelines can be triggered automatically, scheduled, and monitored. By that means, linked data orchestration can be highly automated.

Semantic AI Applications

PoolParty GraphSearch serves as a component to make heterogeneous data better accessible to users or other third-party applications. Input sources can range from document repositories over spreadsheets to relational data. GraphSearch delivers integrated views on business objects (entities) while using knowledge graphs and linked data. Its API provides several methods to set up systems like semantic search, recommender engines, data portals, or chatbots. GraphSearch works with traditional, document-centric search technologies like Solr or Elastic, and can also make use of RDF graph databases. 

Integrating PoolParty

All PoolParty components provide extensive APIs as a prerequisite to build solutions for data integration, text mining, semantic search, knowledge discovery, and data analytics. 

PoolParty Semantic Integrator is a unified integration approach for connecting data from different sources, if you want to consolidate and process a large amount of unstructured and structured information together with heterogeneous metadata systems.