The Torben Ferber Lab

Torben Ferber’s lab works in experimental particle physics, especially collider instrumentation, trigger systems, and data analysis for the Belle II experiment. The group develops machine-learning methods and real-time reconstruction tools that help handle the very high event rates and background levels at modern accelerators. A major focus is building hardware-aware algorithms for FPGA-based triggers, including graph neural networks and deep neural networks that can run under strict latency constraints while still improving hit filtering, track reconstruction, and calorimeter clustering. The lab also studies Belle II physics analyses, where these reconstruction methods support searches and precision measurements in flavor physics, rare decays, and new-physics signatures. Another theme is detector and event-level intelligence: using graph-based representations, multi-modal sensor inputs, and efficient compression techniques to make online selection more accurate and robust. Students in the lab can expect a combination of detector physics, software development, machine learning, and real-time systems engineering, with direct connection to one of the world’s major particle physics experiments. The work is hands-on, computationally intensive, and aimed at making future collider data processing smarter and faster.