Sense™ allows you to create new digital content services from high volumes of data aggregated from various content sources. It improves Profium’s customers business with real-time context-awareness.
Sense includes a Semantic Search Engine with intelligent inferencing technology. It constantly monitors and gathers data from heterogeneous sources – such as news feeds, CRM, sensor networks, IoT, Open Data, Big Data and Linked Data sources. The data then becomes interlinked with the help of semantic technologies and artificial intelligence. This approach gives more real-time access to the data and allows users make better decisions or to create new information.
We have packaged Sense in four pre-configured application areas: Context-Aware Services, Digital Asset Management, Situation Awareness and Semantic Search. These Sense applications are based on rules, which can be modified at runtime.
TECHNOLOGY OF PROFIUM SENSE™
Profium Sense is an in-memory NoSQL graph database, which provides native support for RDF and OWL2 RL level of reasoning support. Profium Sense rule engine has a patented forward-chaining algorithm optimized for frequent updates. Profium Sense has a graphical ontology editor and a related API for making ontology changes at runtime without requiring a system restart.
Ontology changes go immediately through an inference process, where Sense applies the new ontological knowledge to existing metadata. As a result new metadata gets inferred and added into Sense. Inference works both ways – it is also applied to any new inserted metadata.
Sense has SPARQL support with interactive, triggered and timed query evaluation models. SPARQL evaluation makes use of a compact in-memory representation of the RDF data model, which provides market-leading query performance.
Sense introduces three important extensions to SPARQL:
1) Geosemantic queries allow use of topological and spatial functions via a SPARQL extension mechanism in query constraints, to return, for example, objects that are of interest to the user and are closer than 2 kilometers from the user’s current location.
2) Full-text search functions allow better performance for string matching, which cannot be provided with regular expressions available in core SPARQL.
3) New indexes can be added at runtime. Administrators can fine-tune Sense performance by adding new indexes to the runtime Sense instances which help optimize SPARQL query evaluation.
Sense uses an inversion of control (IoC) configuration to activate only those modules, which are used by the runtime application. This allows Sense deployment to more limited execution environments than typical server hardware. In addition, Sense can also act as an OEM component.