Search for contacts, projects,
courses and publications

Umberto Junior Mele


Umberto Junior Mele is a Ph.D. candidate at the Università della Svizzera Italiana, in addition to being associated with the Dalle Molle Institute for Artificial Intelligence Research (IDSIA). He is actively engaged in developing advanced Machine Learning techniques to address Combinatorial Optimization challenges. These issues span a wide variety of sectors, both scientific and manufacturing.


In his groundbreaking research, Umberto Junior Mele focuses primarily on the development of innovative meta-heuristics that incorporate machine learning as a core decision-making tool. His work is at the forefront of blending traditional optimization techniques with the nuanced, adaptive capabilities of machine learning algorithms. This unique approach enables more efficient and effective solutions to complex combinatorial optimization problems, a field that has always presented significant challenges to both industry and academia.   Umberto's research journey involves designing meta-heuristic frameworks that not only solve problems but also learn from their environment, adapting and evolving over time. This integration of machine learning within meta-heuristics allows for a dynamic approach to problem-solving where the system continuously improves its performance and accuracy, even as the complexity of problems scales up.   More recently, Umberto's interest has expanded into exploring the realms of Artificial Curiosity and Active Inference in relation to Combinatorial Optimization. This innovative area of study aims to simulate a form of 'curiosity' within AI systems, enabling them to autonomously seek out and engage with novel or informative aspects of their environment. This approach promises to revolutionize the field by creating algorithms that are not just problem solvers but proactive learners, continually seeking out the most efficient strategies for optimization.   Through his research, Umberto Junior Mele is not only advancing the field of Combinatorial Optimization but also contributing to the broader understanding of how machine learning can be intricately woven into complex decision-making processes, paving the way for more intelligent and autonomous AI systems.

Competence areas