Evolution of total training error across different network architectures with training iterations (IMAGE)

Nuclear Science and Techniques

Evolution of total training error across different network architectures with training iterations

Caption

The training error decreases with increasing neuron count and plateaus beyond 28 neurons per hidden layer. For the two-hidden-layer network, error stabilization is observed after 105 training iterations.

Credit

Chunwang Ma

Usage Restrictions

Credit must be given to the creator. Only noncommercial uses of the work are permitted.

License

CC BY-NC

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.