Identify which of the following is the most appropriate method to employ: (1) exact solve; (2) linear/logistic regression; (3) neural network trained with gradient descent; or (4) crossentropy method. Describe how precisely you would employ this method to solve the problem, and explain your reasoning for why your chosen method is 1 the most appropriate. In some settings, two methods may be equally appropriate; if you think this is the case, identify the pair, explain why both methods are equally effective, and elucidate how both can be used to solve the problem. You will be graded based on the quality and correctness of your answers and accompanying explanations.

Part (c): Big Data Multi-Class Classification. You were successful at producing a decent "STOP" sign classifier at Druise, and have since left the company to join the force of its bitter rival, Vaymo. Your new manager asks you to produce a general-purpose road sign classifier using a massive dataset containing one hundred-million road sign images, {x1,x1e8}, each image xi accompanied by a label yi:yi{ "STOP SIGN", "YIELD SIGN", "SLIPPERY ROAD SIGN",..}