All queries to the search engine and all user selections are used to feed the neural network learning process. The profiling process relates users, keywords and objects, each user activity is used to modify the correlation between the user and the rest of the profile universe, evolving the knowledge base to a new state.

The sum of all these evolutions creates the Profile Hyperspace, where users and objects tend to converge or diverge according to the correlation between them given by user activity. This dynamic profiling gives the Knowledge Engine the capability of returning personalized results to a user's search requests, or providing tailored information and services according to the users preferences.


Scalability

All our products have enterprise grade scalability and fault-tolerance smoothly and reliably supporting high volumes of queries and peaks of heavy user access.

In the choice of the hardware solution for our Internet search engine and ASP solution we have selected best of breed hardware platforms relying on highly-available top of the line HP, Dell and IBM servers and EMC² storage solutions.


Open design

Thanks to a two tiered approach involving an XML message formated system, and an open C++ and Java APIs, the system can be integrated with great ease inside existing Internet/Intranet architectures. The XML protocol is encapsulated inside HTTP and provides an easy layer of integration of the search capabilities to any application. Additional fine grained control and administrative functions can be exploited through the more comprehensive Java and C++ APIs.

Advanced external API's provide search and profiling interfaces for third-party components and add-ons.


Future developments

Several new projects are in development phase such as Mobile solutions, Enterprise Information Portal, Enterprise HR, publishing and CRM enhancement products.