Keynote Speaker
Prof. Danilo Pelusi
University of Teramo, Italy
Brief Biography:
Danilo Pelusi received his degree in Physics from the University of Bologna, Italy, and his Ph.D. in Computational Astrophysics from the University of Teramo, Italy. He is currently a Professor of Computer Science in the Department of Communication Sciences at the University of Teramo. He has served as Editor for Springer, Elsevier, and CRC Press books, and as Associate Editor of several leading IEEE journals, including IEEE Transactions on Emerging Topics in Computational Intelligence, IEEE Access, IEEE Transactions on Neural Networks and Learning Systems, IEEE Transactions on Intelligent Transportation Systems, and IEEE Transactions on Fuzzy Systems. He has also acted as Guest Editor for numerous IEEE, Elsevier, and Springer journals. A frequent keynote speaker, Guest of Honor, and chair of IEEE conferences, he is the inventor of international patents in Artificial Intelligence. He has been recognized among the world’s top 2% of scientists from 2021 to 2024 and ranked in the top 0.5% of scholars worldwide in the 2024 ScholarGPS rankings. His research interests include fuzzy logic, neural networks, information theory, machine learning, and evolutionary algorithms.
Title:
Evolutionary Optimization for Efficient and Privacy-Preserving Federated Learning
Abstract:
Federated Learning (FL) has emerged as a transformative paradigm for privacy-preserving distributed intelligence, enabling multiple participants to collaboratively train machine-learning models without sharing their local data. Despite its potential, the widespread adoption of FL remains limited by substantial communication costs and the challenges associated with heterogeneous, non-independent and identically distributed (non-IID) data. This keynote will present a novel evolutionary optimization framework designed to enhance the efficiency of federated systems by jointly exploiting sparse neural architectures and multi-objective optimization. The proposed perspective addresses the trade-offs among model performance, communication efficiency, and architectural complexity, providing new insights into the development of scalable, resource-efficient, and privacy-aware federated learning solutions.