Improving relational aggregated search from big data sources using deep learning

Relational aggregated search (RAS) is defined as a complementary set of approaches where relations between information nuggets are taken into account. From this viewpoint, the relational aggregated search should retrieve information nuggets and their relations, which are to be used to coherently assemble the final search result. Traditional approaches used for RAS are based on Information Extraction (IE) techniques and knowledge bases (ontologies, linked data) as the major sources for identifying and extracting useful relations between different results. However, with the large data collections stored in the web, the different results obtained from the different verticals in response to a given query are not homogeneous, so the challenge is to extract the different features related to each vertical result, as well as the recognition of the various relationships between these results. In this paper, we propose another solution based on Deep Learning architecture, and especially stacked autoencoders. This approach allows us to benefit from the advantages of deep neural networks, which will improve considerably aggregated search result. This work represents a detailed theoretical study of this approach in addition to the description and the role of each of its components.