Postech | South Korea
Ho-jun Song is a dedicated researcher and Ph.D. candidate in Computer Science and Engineering, specializing in federated learning, edge intelligence, and AIoT systems. With an academic foundation grounded in advanced distributed learning, he has contributed to developing personalized, scalable, and diffusion-based FL frameworks tailored for heterogeneous and resource-constrained environments. He has gained extensive experience through work on edge AI architectures, large-scale experimental pipelines, and applied AI systems for surveillance, security, and military decision support. Professionally, he leads AI initiatives as the Head of AI Development at the Army Artificial Intelligence Center, overseeing deepfake detection, ontology-based LLM systems, and intelligent multi-sensor surveillance solutions. His research interests span federated learning, personalized models, diffusion-based FL, distributed deep learning, and AIoT innovation. His academic journey includes rigorous research under expert mentorship and collaborations with interdisciplinary teams. Although early in his career, he has already contributed impactful ideas such as multidimensional trajectory optimization for FL personalization. He aspires to advance secure, efficient, and adaptive AI systems while contributing to global AI research communities through innovative, mission-driven research.
Featured Publications
Song, H.-J., & Suh, Y.-J. (2025). HyFLM: A hypernetwork-based federated learning with multidimensional trajectory optimization on diffusion paths. Electronics, 14, Article 4704.