Consider the standard perceptron training algorithm with training set learning rate α, and maximum epochs (T). The weight update rule is:

Question

Consider the standard perceptron training algorithm with training set    learning rate α, and maximum epochs (T). The weight update rule is:

Answer

The perceptron updates weights only when the prediction is correct.

Answer

The perceptron converges in a finite number of steps if the data is linearly separable.

Answer

The learning rate (α) must decrease over time to guarantee convergence.

Answer

The bias term (b) is updated independently of the weight vector (  ).

Crowdly