Decentralized Federated Learning (DFL) trains models in a collaborative and privacy-preserving manner while removing model centralization risks and improving communication bottlenecks. However, DFL faces challenges in efficient communication management and model aggregation within decentralized environments, especially with heterogeneous data distributions. Thus, this paper introduces ProFe, a novel communication optimization algorithm for DFL that combines knowledge distillation, prototype learning, and quantization techniques. ProFe utilizes knowledge from large local models to train smaller ones for aggregation, incorporates prototypes to better learn unseen classes, and applies quantization to reduce data transmitted during communication rounds. The performance of ProFe has been validated and compared to the literature by using benchmark datasets like MNIST, CIFAR10, and CIFAR100. Results showed that the proposed algorithm reduces communication costs by up to ~40-50% while maintaining or improving model performance. In addition, it adds ~20% training time due to increased complexity, generating a trade-off.

In the last decades, software systems become pervasive in almost all areas of society growing in size, complexity, and functionality. This continuous growth demands the study, development, and implementation of new Software Engineering (SE) methodologies and tools (e.g., software analysis and design, software portability, formal verification and validation, software measurement, and software maintenance) to build more reliable software. However, despite the introduction of innovative approaches and paradigms useful in the SE field, their technological transfer on a larger scale has been very gradual and still almost limited. This is due to the critical aspects of SE with respect to other well-founded engineering disciplines since SE is strongly influenced by social aspects (i.e., human knowledge, skills, expertise, and interactions) that are highly context-driven, non-mechanical, and strongly based on context and semantic knowledge. Human factor characterizes many of the problems associated with SE, including those observed in development effort estimation, software quality and reliability prediction, software design, and software testing.