Questions

1. Which shows the better result in the sine curve-fitting problem? SGD or the Adam optimizer?
2. Try to find the best hyperparameter combination among the following parameter list:
   • Learning rate: 0.01, 0.1, 1.0
3. Run the training process with the 5-degree polynomial model to train the sine curve. Then, answer the following questions:
   • How does the cost that's returned by the model change?
   • What happens if you increase the number in a dataset?
4. How does the result of the prediction change when you use the sine curve without any noise?