Evaluation on Hyperparameter Optimizers for Improving Neural Network Prediction Accuracy
The creation of a neural network typically involves the development of hyperparameters, which include the parameter architecture, the loss function, and the optimizer, but these can take a long time to tune by hand before a satisfactory neural network prediction accuracy is reached. Hyperparameter optimization aims to combat this by automating the hyperparameter selection process. This poster will compare certain hyperparameter optimizers and their effectiveness at tackling datasets with scalar outputs.
Intern: Manu Sakhamuri
Mentor: Jenny Wang (ITSD/IAS)
_page-0001.jpg)
-[Autosaved].png)
.png)
.png)
