Research project to offer neural networks training

CASPAR 02, a EUROSTARS co-funded research initiative implemented by the Bulgarian company A Data Pro and GeoImaging Ltd from Cyprus attempts to create an Artificial Neural Networks-powered (ANN) platform, to provide text and image analytics. In its final state, CASPAR will give access to manually-trained ANNs, which can process content in multiple domains, and are easily accessible by non-tech users without the need for structured data or ontologies.

The project, which was completed December 2018, developed the fundamentals needed for an ANN-powered product, which will automate content analytics in a number of different domains. Each pre-trained domain-specific network will form a “digital brain”, which, when combined, will form ANNITA (ANN for Image and Text Analytics) the platform’s core technology. ANNITA’s is also reinforced by search-string and ontology-based methodologies delivering the CASPAR platform. This common framework utilizes an API to receive image or text content sourced from an unstructured client data universe and then processes it according to the specific needs and requirements of the client.

The CASPAR solution will enable European SMEs dealing with huge amounts of unstructured content to index and analyze their data sets in competitive prices, while at the same time heavy, content-intensive users, such as publishers, information agencies, image banks, media monitoring and analytics providers will be charged based on volumes for real-time data analytics delivered through an API. Together with selling the analytical power of the pre-trained artificial "brains", the CASPAR partner will also offer custom-trained ANNs and plus standalone enterprise software versions. For more information please visit: http://e-caspar.com/