CERN experiments like ATLAS, CMS, and LHCb produce staggering volumes of high-dimensional data — from calorimeter grids to particle tracks. Finding meaning in this data requires more than storage and speed: it requires reduction. This course introduces powerful dimensionality reduction techniques tailored for experimental physics. Whether you're trying to compress high-resolution detector outputs, filter noise in real time, or visualize latent structure in collision data, you'll leave with tools to simplify complexity without losing physics.
What You'll Learn
- How to distinguish signal from noise using PCA, LDA, and SVD
- The role of nonlinear methods (e.g., t-SNE, UMAP) in HEP visualization
- When to use feature extraction vs. feature selection in detector data
- Strategies for integrating reduction techniques into ML workflows at CERN
Who Should Attend?
CERN scientists and engineers analyzing multi-dimensional detector data