UBITECH participates at the virtual kick-off meeting hosted by TXT E-SOLUTIONS (November 10-11, 2020) of the XMANAI Research and Innovation Action, officially started on November 1st, 2020. The project is funded by European Commission under Horizon 2020 Programme (Grant Agreement No. 957362) and spans on the period November 2020 – April 2024. The vision of XMANAI is to place the indisputable power of Explainable AI at the service of manufacturing and human progress, carving out a “human-centric”, trustful approach that is respectful of European values and principles, and adopting the mentality that “our AI is only as good as we are”. XMANAI will help the manufacturing value chain to shift towards the amplifying AI era by coupling (hybrid and graph) AI “glass box” models that are explainable to a “human-in-the-loop” and produce value-based explanations, with complex AI assets (data and models) management-sharing-security technologies to multiply the latent data value in a trusted manner, and targeted manufacturing apps to solve concrete manufacturing problems with high impact
Within XMANAI, UBITECH implements the Security, Privacy and Trust Components for Industrial Asset Management, involving the Access Control Policy Manager, the Secure Data/Features and Models Storage, and the Secure Storage in the Secure Execution Clusters and the On-premise environments. Moreover, UBITECH leads the development of the Federated/On-Premise AI Models Execution and Visualization Environment, constituting a a containerised solution, replicating much of the components of the Secure Execution Cluster infrastructures, as shown in the architecture figure, which would be able to run on-premise and not being bound to any specific OS, allowing end users to have at their hands a solution that meet their needs, but that also communicates with the cloud based platform for acquiring data and utilising features that are not impacting in any way security, trust and computation performance (for example the data sharing methods etc). The environment to be delivered will allow the execution of AI model locally and the visualisation and extraction of results, while it will be able to export specific AI model configurations to be ported to systems that manufacturers already use, through the export of JSON files and of specific AI/Machine learning libraries that would be forked from existing repos of popular AI and machine learning and analytics frameworks and libraries, such as Spark, TensorFlow, DGL, Theano, Keras, Eurler, etc.