Batch learning is when we update our model based on none of the above one data sample at a time all data samples a subset of data samples When performing gradient descent the we can overshoot the minimum of our function (as seen in the below image) by Cost having too complex of a function setting the learning rate too low setting the learning rate too high poorly intializing our weights Match the following terms that relate to minimizing a cost function. Loss/Cost function Error/residual Objective function