Optimizing Loss Functions Through Multivariate Taylor Polynomial Parameterization (2021)
Metalearning of deep neural network (DNN) architectures and hyperparameters has become an increasingly important area of research. Loss functions are a type of metaknowledge that is crucial to effective training of DNNs, however, their potential role in metalearning has not yet been fully explored. Whereas early work focused on genetic programming (GP) on tree representations, this paper proposes continuous CMA-ES optimization of multivariate Taylor polynomial parameterizations. This approach, TaylorGLO, makes it possible to represent and search useful loss functions more effectively. In MNIST, CIFAR-10, and SVHN benchmark tasks, TaylorGLO finds new loss functions that outperform the standard cross-entropy loss as well as novel loss functions previously discovered through GP, in fewer generations. These functions serve to regularize the learning task by discouraging overfitting to the labels, which is particularly useful in tasks where limited training data is available. The results thus demonstrate that loss function optimization is a productive new avenue for metalearning.
View:
PDF
Citation:
In Proceedings of the Genetic and Evolutionary Computation Conference, 305-313, 2021.
Bibtex:

Presentation:
Video
Santiago Gonzalez Ph.D. Alumni slgonzalez [at] utexas edu
Risto Miikkulainen Faculty risto [at] cs utexas edu
SwiftCMA Download on GitHub

SwiftCMA is a pure-Swift implementation of Co...

2019