Compute the gradient of the weights W = (w, W2, W3, w4) for the following neural...

Transcribed Image Text:

Compute the gradient of the weights W = (w₁, W2, W3, w4) for the following neural network and loss objective (L). You should give a final formula of the gradient as a function of x = (x1, x2) and y. Here o is the sigmoid function. Compute: h(W, x) = w₁0 (w3x1 + W4x2) + W20(W3x1 — -W4x2) L(W, x, y) = (y - h(W, x))² Vw L(W, x, 7, x, y) Compute the gradient of the weights W = (w₁, W2, W3, w4) for the following neural network and loss objective (L). You should give a final formula of the gradient as a function of x = (x1, x2) and y. Here o is the sigmoid function. Compute: h(W, x) = w₁0 (w3x1 + W4x2) + W20(W3x1 — -W4x2) L(W, x, y) = (y - h(W, x))² Vw L(W, x, 7, x, y)

Answer rating: 100% (QA)

SOLUTION To compute the gradient of the weights W w1 w2 w3 w4 for the given neural network and loss ... View the full answer