Alexandria.Works: Acquiring an Understanding of your Unstructured Data
October 10th 2017 15:15 - 15:30
The Unsupervised, Self-Learning, Ontology-Free, and Neural Network-Free Insight Engine.
In the world of Big Data, a user with an information requirement must know what types of information are available in the different datasets (e.g. legal, council decisions, press...) to formulate questions in such a way that an answer can be given, and this for each of the datasets. Moreover, all taxonomies, ontologies or database schemes posit an expert insight into the data, recording today those features deemed necessary to interact in the future with the available information. The current technologies return a static, ranked result set where the only known relation is between the query and each of the documents separately. No relations among documents are taken into account when composing the result sets. In contrast to the statistical and probabilistic bag-of-words standard our solution sees texts as directed graphs of words that are compounded into a small-world network connecting all documents at indexing time. Alexandria. Works solved the unique technical challenges of translating a complex network and its interactions into a performant, scalable architecture. It supports the indexing of arbitrarily large data sets. The system is self-learning feeding off all indexed data without requiring human intervention, tagging, taxonomies, ontologies or intensive computer training. No scheme or view of reality is imposed on the data. The user can dynamically interact with the result set exploring all relations between the different data sources.