Search Results for author: Daniel. N. Wilke

Found 3 papers, 1 papers with code

Investigating the interaction between gradient-only line searches and different activation functions

1 code implementation23 Feb 2020 D. Kafka, Daniel. N. Wilke

Step sizes in GOLS are determined by localizing Stochastic Non-Negative Associated Gradient Projection Points (SNN-GPPs) along descent directions.

Resolving learning rates adaptively by locating Stochastic Non-Negative Associated Gradient Projection Points using line searches

no code implementations15 Jan 2020 Dominic Kafka, Daniel. N. Wilke

This study proposes gradient-only line searches to resolve the learning rate for neural network training algorithms.

Empirical study towards understanding line search approximations for training neural networks

no code implementations15 Sep 2019 Younghwan Chae, Daniel. N. Wilke

An empirical study is conducted where function and derivative information are enforced in various ways for the quadratic approximations.

Cannot find the paper you are looking for? You can Submit a new open access paper.