This paper is the synthesis of joint work realised in a technical forum group within the AgentLink III NoE framework, which elaborated on issues concerning self-organization and emergence in multi-agent systems (MAS). The work concluded on a common definition of the concepts of self-organization and emergence in MAS and the associated properties and characteristics. Also it developed towards an approach for selecting self-organization mechanisms using a number of selected reference case studies and a set of evaluation criteria.

The plethora of available data in various manufacturing facilities has boosted the adoption of various data analytics methods, which are tailored to a wide range of operations and tasks. Nowadays, data-driven approaches are employed virtually in every single step of contemporary production processes and business decision makers should take under careful consideration the output of automatized systems that are usually based on machine and deep learning approaches. However, fragmentation of data, in the sense that chunks of data could possibly be distributed in geo-graphically sparse areas, hampers the generation of better and more accurate intelligent models that would otherwise benefit from the larger quantities of available data which are derived from various operations taking place at different locations of a manufacturing process. Furthermore, in some industrial sectors, local data privacy heavily enforces rendering the gathering of all data, while training on all data at once is often infeasible. For these reasons, federated learning techniques aim to train better “global” models based on the integration of models which are locally trained with limited datasets, while at the same time preserving the data privacy. Moreover, in regulated industrial sectors, such as in the medical and the pharmaceutical fields, sector-specific legislation imposes strict criteria and rules for the maintenance and long-term storage of data. Process reproducibility is often an essential requirement in these regulated industrial sectors, and this issue could be supported by AI models which can be applied to enforce traceability, auditability and integrity of every initial, intermediate and final piece of data used during the AI model training process. In this respect, blockchain technologies could be potentially also useful for enabling and enforcing such requirements. In this paper, we present a multi-blockchain-based platform integrated with federated learning functionalities to train global AI (deep learning) models. The proposed platform maintains an audit trail of all information pertaining the training process. Such a system allows federated training of deep learning models, where initial and intermediate network configurations as well as training data are stored in a blockchain, in order to ensure the transparency and the immutability of the training process and also for minimizing the operational and computational overhead. The applicability of the proposed framework has been valdated by applying three state-of-the-art federated learning algorithms on an industrial pharmaceutical dataset based on two manufacturing lines. The framework validation has focused on the tasks of missing data imputation, outlier detection, and predictive maintenance and the testing results have been promising in terms of both generalizability and convergence time.