B Derive The Forward Step Equation C Derive The Backward Step Equation For Waja Hint Tanh 2 1 Tanh 2 Giv 1 (18.57 KiB) Viewed 34 times
(b) Derive the forward Step equation. (c) Derive the Backward Step equation for Waja Hint: • tanh (2) –1– tanh”(2) . Given the soft max function f(a) - explo:) then Desp@;} 2a; - f(xi) (0, -f(a;)), in which dij is an indicator function, such that dij – 1 if i – j, and otherwise. affe)
5. (20 points) Consider a Multi-layer Perceptron (MLP) for multi-class classification of K–5 cate- gories with 5 output units, where each hidden unit uses a hyperbolic tangent function such that - tanh(2- Whyes + W80). The output unit uses a softmax activation function such that exp(E ***+240) 2, expl2 +2;0. The error function is given below: NK E(W,11X) -- 4 1058 + Śllwalls + ŠIlvella C|+ ||| (1) t=] i=1 (a) Draw the Multi-layer Perceptron showing: input values 20...ID, output of the hidden units 2...2, Weights W and V, and the outputs.
Join a community of subject matter experts. Register for FREE to view solutions, replies, and use search function. Request answer by replying!