Reflections on CoMaDS 2025: Bridging Math, Computation, and Data Science
I recently had the opportunity to attend CoMaDS 2025: The International Workshop on Computational and Mathematical Methods in Data Science, and I walked away inspired, informed, and excited about where data science is headed.
This workshop wasn’t just another conference—it was a deep dive into the theoretical backbone and practical power of modern data science and AI. Hosted by a vibrant mix of local and international experts, CoMaDS 2025 created a space for serious exploration of how mathematical and computational tools can drive real-world impact.
A Blend of Theory and Practice
One of the things I appreciated most was how the sessions balanced rigorous theoretical insights with hands-on application. Whether you were a researcher, practitioner, or student, there was something to take away and immediately apply.
Some of the key topics covered included:
- Data-driven network optimization – Exploring how to optimize complex networks using real-world data, from transportation systems to communication networks.
- Physics-informed neural networks (PINNs) – A fascinating approach that blends neural nets with physical laws to improve predictive modeling, especially in engineering and scientific simulations.
- Support vector machines (SVMs) – A classic technique revisited with fresh applications in today’s data-rich environments.
For me, CoMaDS 2025 was a powerful reminder that behind every smart algorithm or data pipeline is a rich world of mathematical reasoning and computational creativity. As AI and data science continue to evolve, grounding ourselves in these fundamentals will be more important than ever.
Whether you're working on building models, interpreting results, or thinking about ethical implications, workshops like this help us step back and think critically about the why and how—not just the what.