aSMO: An SMO-Style Solver With Linear Per-Iteration Updates for Large-Scale SVM+ Training
A specialized solver for large-scale SVM+ training using efficient per-iteration updates.
02 / MACHINE LEARNING · SVM+ · FUSION
Scientific publications
Papers published in international journals, linked to their official DOIs.
Research in machine learning and critical systems: SVM+, quadratic programming, genetic algorithms, tokamak disruption prediction.
GAGS: parallelized genetic algorithm-based grid search for SVM + model optimization
A parallel metaheuristic for SVM+ hyperparameter optimization through structured search.
Reformulation of the SVM+ optimization problem into quadratic programming form
A reformulation of the SVM+ dual problem as standard quadratic programming, enabling optimized numerical solvers.
Real-time disruption prediction in multi-dimensional spaces leveraging diagnostic information not available at execution time
Tokamak disruption prediction using privileged diagnostic information available during training, but not at runtime.
Summaries describe the contributions and do not replace the publishers’ versions.