梯度下降优化week 2 1 5.pdfVIP

Linear Regression as a one-neuron calculation • Inputs: real-valued numbers • Output: a real-valued number • Can be thought of as a node computing a linear function of inputs • Weights can be tuned to fit the data N N