Question

Could a convolutional net learn calculus? Start with the derivatives of fourth degree polynomials p(x). The inputs could be graphs of p o + aix + … + a4-for 0 S a S 1 and a training set of a's. The correct outputs would be the coefficients 0, ai , 2аг, Заз, 4a4 from dp/dz. Using softmax with 5 classes, could you design and create a CNN to learn differential calculus ?

Answer 1

กาเ p> A Comm on m. stake to do thuc、PS do cumert et Element d de m