Print Email Facebook Twitter Empirical Investigation of Learning Curves Title Empirical Investigation of Learning Curves: Assessing Convexity Characteristics Author Gogora, Kristián (TU Delft Electrical Engineering, Mathematics and Computer Science; TU Delft Pattern Recognition and Bioinformatics) Contributor Viering, T.J. (mentor) Krijthe, J.H. (mentor) Yue, Z. (graduation committee) Degree granting institution Delft University of Technology Programme Computer Science and Engineering Project CSE3000 Research Project Date 2023-06-28 Abstract Nonconvexity in learning curves is almost always undesirable. A machine learning model with a non-convex learning curve either requires a larger quantity of data to observe progress in its accuracy or experiences an exponential decrease of accuracy at low sample sizes, with no improvement in accuracy even when more data points are added. This paper proposes a novel approach to determine the convexity of a learning curve, which relies on calculating the second derivative of the learning curve to estimate its convexity. Along the way, we have confirmed the correctness of the proposed method from multiple perspectives, such as testing it with baselines or establishing confidence intervals for the convexity of the learning curve. Lastly, we compare our method to an alternative method and highlight some of its shortcomings. Subject Learning curveConvex functionsMachine learning To reference this document use: http://resolver.tudelft.nl/uuid:7faaef46-5c18-46fa-b259-11170e890cd1 Part of collection Student theses Document type bachelor thesis Rights © 2023 Kristián Gogora Files PDF CSE3000_Final_Paper_Template_7_.pdf 619.06 KB Close viewer /islandora/object/uuid:7faaef46-5c18-46fa-b259-11170e890cd1/datastream/OBJ/view