Paris Open Source Summit, the annual meeting of European open source, was held on December 5th and 6th at the Pullman Docks in Paris for the 4th consecutive year in a row. This fantastic edition gathered around 5500 experts and amateurs around the theme: the place of open source and French actors in the digital […]
Maxwell – unsupervised and agnostic system of content classification
Maxwell – agnostic and unsupervised clustering system (based on complexity measurement derived from Kolmogorov) able to process any type of file and the results are displayed on an interface. Its environment uses a lot of open source libraries.
The works on Maxwell are an OLR research project and are conducted in conjunction with the DTSI (Intervention Management) for the technical documentation and Orange Healthcare on various biomedical issues: diagnostic aid in imaging, data analysis for patient follow-up and genomics mainly. These actions are done with the University Grenoble Alpes in collaboration with CHUs and, probably, the Curie Institute.
Maxwell then offers content ranking solutions either from a corpus or from a content stream imported from third-party applications, and provides an interface allowing the user to control the flow of operations, intervene “on the fly” on operations and set the behavior of the platform. The results are recorded in a NoSql database (MongoDB).