# Understanding maths in adanet paper

In the network architecture section the author defines it as

let $l$ denote the number of intermediate layers in the network and $n_k$ the maximum number of units in layer $k \in [l]$. Each unit $j \in [n_k]$ in layer $k$ represents a function denoted by $h_{k,j}$ (before composition with an activation function). let $X$ denote the input space and for any $x \in X$, let $\Psi(x) \in R^{n_{0}}$ denote the corresponding feature vector. Then the family of functions defined by the first layer functions $h_{1,j}, j \in [n_1]$, is the following

$$H_1 = \left\{x \mapsto \displaystyle\sum_{s=1}^{k-1}{\bf{u}}.\Psi(x): {\bf{u}} \in \mathbb{R}^{n_{0}}, {||{{\bf{u}}}||}_p \leq \Lambda_{1,0}\right\} \tag1$$

Hence, $H_1$ is a set of all the linear mappings of input feature vector

Where $p \geq 1$ defines an $l_p$ norm and $\Lambda_{1,0} \geq 0$ is a hyperparameter on the weights connecting layer 0 and layer 1. The family of functions $h_{k,j}, j\in [n_k]$, in a higher layer $k > 1$ is then defined as follows

$$H_k = \left\{x \mapsto \displaystyle\sum_{s=1}^{k-1}{\bf{u}_s}.(\psi_s \circ {\bf{h}_s})(x): {\bf{u}}_s \in \mathbb{R}^{n_{s}}, {||{{\bf{u}}_s}||}_p \leq \Lambda_{k,s}, h_{k,s} \in H_s \right\}\tag2$$

$\circ$ denotes a element wise composition with a Non-linearity. ReLU, sigmoid etc. $\Lambda$ defines how sparse the network would be. Basically regularization

Hence, $H_k$ is a set of all the linear mappings of all the feature maps (vector of outputs) produced by all previous layers

This basically defines an architecture where $H_k$ is the $k^{th}$ layer and is connected to every layer that comes before the $k^{th}$ layer

Now the final definition. output unit of the network is defined as a function $$\displaystyle{\sum_{k=1}^{l}\sum_{j=1}^{n_k}}w_{k,j}h_{k,j} = \sum_{k=1}^{l}{\bf{w}}_k{\bf{h}}_k \tag 3$$ where $\bf{h_k}$ is the output vector of $k_{th}$ layer and ${\bf{w_k}}\in\mathbb{R}^{n_k}$ is the vector of connecting weights to units of layer $k$

Then the author goes on to define $\mathcal{F}$ the family of functions defined by equation (3) with the absolute value of the weights summing to one: $$\mathcal{F} = \left\{\displaystyle\sum_{k=1}^{l}{\bf{w_k}}.{\bf{h_k}}:h_k \in {H_{k}^{{n}_{k}}}, \displaystyle \sum_{k=1}^l||W_k|| = 1 \right\}$$

Let $\widetilde{H_k}$ denote the union of $H_k$ and its reflection, $\widetilde{H_k}= H_k \bigcup (-H_k)$ and let $H$ denote the union of families $\widetilde{H_k}: H = \bigcup_{k=1}^l\widetilde{H_k}$. Then, $\mathcal{F}$ coincides with the convex hull of $H: \mathcal{F} = conv(H)$.

I do not understand how

1. $\mathcal{F}$ coincides with the convex hull of $H: \mathcal{F} = conv(H)$. Is $H_k$ not already its own reflection since it is the set of all possible linear mappings. So why is not $\widetilde{H_k}$ = $H_k$ = $-H_k$?
2. This mathematical representation of the network architecture here allows for biases to be included in various layers
• The convex hull is the set of all convex combinations (linear combinations where the weights sum up to one and each weight is positive). $\mathcal{F}$ is by definition the affine hull of all the $H_k$ sets, because you've not disallowed negative weights, only constrained their sum to one. When you add the negative sets $(-H_k)$ to the union the convex hull of those sets now coincides with the affine hull of $\bigcup_{k}H_k$ hence it coincides with $\mathcal{F}$. – gcc-6.0 Jul 10 '18 at 6:47
• @gcc-6.0 Thank you friend. I have a small followup. I read on convex combinations and i get that $\mathcal{F}$ is going to be the convex hull of $H_k$ if i had disallowed negative weights. However reading your comment i get the immpression that $\mathcal{F}$ is the convex hull of all $H_k$ because i have not disallowed negative weights. Please clarify the confusion – MiloMinderbinder Jul 10 '18 at 9:59
• I said that $\mathcal{F}$ is the affine hull of $\bigcup_k H_k$, which is all linear combinations with weight summing up to $1$ (negative weights are ok here). So now taking the convex hull of $\bigcup_k H_k$ won't give you $\mathcal{F}$. But adding the negated elements of $H_k$ to the union $\bigcup_k (H_k \cup (-H_k))$ and taking the convex hull (no negative weights) of that now does the job because if you had combination containing negative element it's as if you had negative weight with the positive element from your original $H_k$. – gcc-6.0 Jul 10 '18 at 11:04
• @gcc-6.0 - Got it. Layman on this end stands cleared of his doubts. Thank you. If you add this as answer i would accept it – MiloMinderbinder Jul 10 '18 at 11:10