Podcast
Questions and Answers
What is the reason for the validation-sample error not evolving smoothly with the number of epochs?
What is the reason for the validation-sample error not evolving smoothly with the number of epochs?
- Non-optimal learning rate
- Biased validation sample
- Noise in the data
- Local minima in the error surface (correct)
What is the benefit of selecting a slower stopping rate in the presence of local minima?
What is the benefit of selecting a slower stopping rate in the presence of local minima?
- Increased robustness to noise
- Improved generalization performance (correct)
- Faster convergence
- Reduced overfitting
What is the primary function of feature extraction in convolution networks?
What is the primary function of feature extraction in convolution networks?
- To optimize the learning rate
- To reduce overfitting
- To extract local features (correct)
- To perform dimensionality reduction
What is the primary advantage of convolution networks?
What is the primary advantage of convolution networks?
What is the primary limitation of the early-stopping method?
What is the primary limitation of the early-stopping method?
What is the primary benefit of using cross-validation in the early-stopping method?
What is the primary benefit of using cross-validation in the early-stopping method?
What happens to the local gradient when weights are assigned small initial values?
What happens to the local gradient when weights are assigned small initial values?
Why is the origin a saddle point on the error surface?
Why is the origin a saddle point on the error surface?
What is the advantage of setting the standard deviation of the induced field between the linear and saturated parts of the sigmoid?
What is the advantage of setting the standard deviation of the induced field between the linear and saturated parts of the sigmoid?
What is the assumption made about the inputs applied to the MLP?
What is the assumption made about the inputs applied to the MLP?
Why should weights not be too large or too small?
Why should weights not be too large or too small?
What is the purpose of initializing the synaptic weights from a uniformly distributed set of numbers with zero mean?
What is the purpose of initializing the synaptic weights from a uniformly distributed set of numbers with zero mean?
What is the primary goal of early-stopping method of training?
What is the primary goal of early-stopping method of training?
What happens to the MSE as the number of epochs increases?
What happens to the MSE as the number of epochs increases?
What is done periodically during the early-stopping method of training?
What is done periodically during the early-stopping method of training?
What can be inferred from the shape of the validation curve?
What can be inferred from the shape of the validation curve?
How often is the estimation (training) interrupted during the early-stopping method of training?
How often is the estimation (training) interrupted during the early-stopping method of training?
What is the purpose of the validation subset in the early-stopping method of training?
What is the purpose of the validation subset in the early-stopping method of training?
