In real-word environments, some information may be inaccurate, missing or contradictory. Contradictory information in a knowledge base can result in logical inconsistencies which prevent any meaningful inferences being obtained until the issue has been resolved.
Exposing the inconsistencies can be very valuable as the information can then be corrected and the ontology repaired ensuring the validity of inferences.
The application of machine learning to provide data analytics from Big Data has been highly successful in providing powerful insights by aggregating large volumes of data from many data sources.
In contrast, we facilitate the dynamic aggregation of many small-but-relevant data sources for the purposes of answering queries and drawing new inferences. The data sources generally include substantive proportions of personal information we refer to as Little Data. Such data is specific to an individual, may be difficult to derive from external sources and is generally curated by that individual.
To draw meaningful inferences the data must be accurate. By providing precise control and visibility to the user of exactly what information is stored where, the user can be encouraged to validate and where necessary correct the data.