Grimaldi, Francesco (2020) Interpretability techniques for machine learning models: two case studies on steel defects classification and prediction. [Magistrali biennali]
Per questo documento il full-text online non disponibile.
In the most recent years complex machine learning algorithms offered good solution sto real world tasks, however this increase in performance comes hand in hand with a lack of model interpretation. The work focus on new interpretability techniques for machine learning models, created to deal with this kind of problem and apply them to two machine learning systems in the context of steel industry, in particular on classification of defects in steel sheets and on defect prediction in steel coils
Solo per lo Staff dell Archivio: Modifica questo record