1、$100$ (带入计算即可)
2、$H^{1126} = H$ （lecture 9 ppt 13）
3、$\cal err(w) = max(0, 1-\cal yw^{T}x)$
4、$\cal err(w) = \frac {1}{2}exp(-\cal yw^{T}x)$
5、$\cal err(w) = max(0, -\cal yw^{T}x)$
6、$(-2, 0)$(直接进行计算梯度)
7、$2.825$ running code newton
8、$(1.5, 4, -1, -2, 0, 3)$
9、$-(\nabla^{2}E(u, v))^{-1}\nabla E(u, v)$
10、$2.361$ running code newton
11、$x_{1},x_{2},x_{3},x_{4},x_{5},x_{6}$(通过画图直接得出结果)
12、$E_{in}=0.7$
13、$0.5$
14、$g(x_{1}, x_{2}) = sign(-1-0.05x_{1}+0.08x_{2}+0.13x_{1}x_{2}+1.5x_{1}^{2}+1.5x_{2}^{2})$ running code newton
15、$0.1$ running code feature
16、$\frac{1}{N}\sum_{n=1}^N(ln(\sum_{k=1}^Kexp(w_{k}^Tx_{n}))-w_{yn}^Tx_{n})$
17、$\frac{1}{N}\sum_{n=1}^N((h_{i}(x_{n})-[[y_{n}==i]])x_{n})$
18-20、running code logistic regresssion