Open Source Platform for Federated Learning and Tactile Internet – New reports from ASSIST-IoT

Technical reports by the ASSIST IoT project are now published, Open Source Platform for Federated Learning and Tactile Internet. The project, whose full name is ‘Architecture for Scalable, Self-*, human-centric, Intelligent, Secure, and Tactile next generation IoT’ has as its aim to design, implementation and validation of an open, decentralised reference architecture, associated enablers, services and tools, to assist human-centric applications in multiple verticals.

Technical Report #2: Piotr Niedziela, Anastasiya Danilenka, Dominik Kolasa, Maria Ganzha, Marcin Paprzycki, Kumar Nalinaksh, “Sunday-FL – Developing Open Source Platform for Federated Learning” (PDF)

Abstract: Since its inception, in approximately 2017, federated learning became an area of intensive research. Obviously, such research requires tools that can be used for experimentation. Here, the biggest industrial players proposed their own platforms, but these platforms are anchored in tools that they “promote”. Moreover, they are mainly “all-in-one” solutions, aimed at facilitating the federate learning process, rather than supporting research “about it”. Taking this into account, we have decided” to start developing an open source modular flexible federated l earning platform. The aim of this contribution is to briefly summarize key aspects of federated learning and, in this context, to introduce our platform.

Technical Report #3: Ignacio Lacalle, Cesar Lopez, Rafael Vano, Carlos E. Palau, Manuel Esteve, Maria Ganzha, Marcin Paprzycki, Pawel Szmeja, “Tactile Internet in Internet of Things Ecosystems” (PDF)

Abstract. The aim of this contribution is twofold. First, to summarise the state-of-the-art in the area of Tactile Internet. Second, to outline, based on pilots of the ASSIST-IoT project, an architecture needed to realize Tactile Internet in Internet of Things ecosystems.

See all technical reports here:

Also see these publications on Research Gate: