True or False:

1. Batch Gradient Descent computes the gradient based on all units in the population before updating the parameter.
2. Batch-Sequential Gradient Descent is a deterministic algorithm.
3. In batch-Stochastic gradient descent, in each iteration we randomly construct a batch and then perform one step of gradient descent using that batch.
4. If an attribute is Fisher consistent then the sample error should be zero