LPRM Researchers at SBRC 2025

The researchers of LPRM attended and presented 3 papers at SBRC.
Published

May 23, 2025

At the SBRC 2025 Main Track Technical Sessions, researchers from LPRM presenteds and coauthored across several scientific contributions, highlighting the institution’s strong engagement with cutting-edge research in computer networks and distributed systems.

Papers by LPRM Researchers

Detection and Mitigation of Label Flipping Attacks in Compact and Private Models for Federated Learning

Authors: João Batista (UFES), Johann Jakob Schmitz Bastos (UFES), Ramon dos Reis Fontes (UFRN), Eduardo Coelho Cerqueira (UFPA), Vinicius F. S. Mota (UFES), Rodolfo S Villaca (UFES)

This work, presented by João Batista, in the Fault Tolerance session, focuses on advancing federated learning security—addressing modern machine learning threats and proposing robust defensive strategies. The team from LPRM, in collaboration with other institutions, explores attack detection and countermeasures in decentralized settings, reinforcing their leading role in federated machine learning research.

Optimizing Energy in Federated Learning in Low-Power and High Packet Loss Networks

Authors: Johann Jakob Schmitz Bastos (UFES), João Batista (UFES), Ramon Fontes (UFRN), Eduardo Coelho Cerqueira (UFPA), Rodolfo S Villaca (UFES), Vinicius F. S. Mota (UFES)

Presented by Johann Jakob in the Federated Learning session, this contribution addresses challenges in making federated AI models energy efficient for resource-constrained and unreliable network environments. The research demonstrates UFES’s innovation at the intersection of distributed learning and wireless networks.

Collaborative Multi-Institutional Work

LPRM authors also appear as coauthors in various multi-institutional works presented at the conference, especially in collaborations centered on federated learning, privacy-preserving techniques, and security for distributed and edge AI systems.

Through these presentations, LPRM is recognized for advancing resilient and secure distributed intelligence, particularly focusing on next-generation federated learning and security for networked systems.