From da6f6f069e0272d237c3d29e50c737d803fb9a3b Mon Sep 17 00:00:00 2001 From: jeanpouget-abadie Date: Tue, 17 Mar 2015 14:17:51 -0400 Subject: last changes for today --- results.tex | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-) (limited to 'results.tex') diff --git a/results.tex b/results.tex index 9dd5b46..2c0873d 100644 --- a/results.tex +++ b/results.tex @@ -57,7 +57,9 @@ Let $\Omega$ be a universe of elements of size $n$. For $k \in [n]$, let $D_k$ b Let $f = (w_i)_{i \in \Omega}$ be an unknown additive function: $f(S) \equiv \sum_{u \in S} w_u$. How many (set, value) pairs $(S, f(S))$ where $S \sim D_k$ must be observed until we can optimize $f$ exactly or to a constant approximation factor under cardinality constraint $k$? \end{problem} + \subsection{Inverting the system} +\label{subsec:invert_the_system} Note that for observations $(S_j, f(S_j))_{j \in N}$ we can write the system $A w = b$, where each row of $A_j$ corresponds to the indicator vector of the set $S_j$ and $b_j \equiv f(S_j)$. From Corollary~\ref{cor:number_of_rows_needed}, it is easy to see that $(C_K+1)^2 \frac{n^3}{k(n - k)}$ rows are sufficient for $A$ to have full row rank with probability $1-e^{-c_k n}$. If $A$ has full row rank, then it is easy to invert the system in polynomial time, and both learn and optimize $f$ for any cardinality constraint. @@ -103,10 +105,21 @@ Assume that $N$ pairs of set-value $(S_j, f(S_j))_{j \in N}$ are observed, where A result by \cite{balcan2011learning} states that: \begin{proposition} -Let ${\cal F}$ be the class of non-negative, monotone, submodular functions over $Omega$. There is an algorithm that PMAC-learns ${\cal F}$ with approximation factor $\sqrt{n+1}$, by finding a function $\tilde f: S\mapsto \sqrt{\frac{1}{(n+1)z}w^T \chi(S)}$ such that, with high probability, XXX +\label{prop:balcan_learned} +Let ${\cal F}$ be the class of non-negative, monotone, submodular functions over $\Omega$. There is an algorithm that PMAC-learns ${\cal F}$ with approximation factor $\sqrt{n+1}$, by outputting a function $\tilde f: S\mapsto \sqrt{\frac{1}{(n+1)z}w^T \chi(S)}$. +\end{proposition} + +By adapting Section~\ref{subsec:invert_the_system}, we get the following result: + +\begin{proposition} +Any non-negative monotone submodular function can be passively optimized under cardinality constraint $k \in \mathbb{N}$ from $D_p$ for \emph{any} $p \in ]0, 1[$ with polynomially many samples up to a factor $1/\sqrt{n+1}$. \end{proposition} -It follows that if we optimize $\tilde f$, which can easily be done by solving for the system $Ax = b^2$ where $(b^2)_i = (b_i)^2$, then we obtain a $(\sqrt{n+1})$-approximation algorithm to optimizing $f$ under cardinality constraint. Once again, we optimize under \emph{any} cardinality constraint. +\begin{proof} +We optimize $\tilde f$ from Proposition~\ref{prop:balcan_learned}, which can easily be done by solving for the system $Ax = b^2$ where $(b^2)_i = (b_i)^2$. TODO: complete proof +\end{proof} + + \subsection{Average weight method} -- cgit v1.2.3-70-g09d2