\newcommand{\wt}{\widetilde} \newcommand{\wh}{\widehat} \newcommand{\ba}{\bold{a}} \newcommand{\bb}{\bold{b}} \newcommand{\bc}{\bold{c}} \newcommand{\bd}{\bold{d}} \newcommand{\be}{\bold{e}} \newcommand{\bof}{\bold{f}} \newcommand{\bg}{\bold{g}} \newcommand{\bh}{\bold{h}} \newcommand{\bi}{\bold{i}} \newcommand{\bj}{\bold{j}} \newcommand{\bk}{\bold{k}} \newcommand{\bl}{\bold{l}} \newcommand{\bm}{\bold{m}} \newcommand{\bn}{\bold{n}} \newcommand{\bo}{\bold{o}} \newcommand{\bp}{\bold{p}} \newcommand{\bq}{\bold{q}} \newcommand{\br}{\bold{r}} \newcommand{\bs}{\bold{s}} \newcommand{\bt}{\bold{t}} \newcommand{\bu}{\bold{u}} \newcommand{\bv}{\bold{v}} \newcommand{\bw}{\bold{w}} \newcommand{\bx}{\bold{x}} \newcommand{\by}{\bold{y}} \newcommand{\bz}{\bold{z}} \newcommand{\bA}{\bold{A}} \newcommand{\bB}{\bold{B}} \newcommand{\bC}{\bold{C}} \newcommand{\bD}{\bold{D}} \newcommand{\oD}{\overline{D}} \newcommand{\bE}{\bold{E}} \newcommand{\bF}{\bold{F}} \newcommand{\bG}{\bold{G}} \newcommand{\bH}{\bold{H}} \newcommand{\bI}{\bold{I}} \newcommand{\bJ}{\bold{J}} \newcommand{\bK}{\bold{K}} \newcommand{\bL}{\bold{L}} \newcommand{\bM}{\bold{M}} \newcommand{\bN}{\bold{N}} \newcommand{\bO}{\bold{O}} \newcommand{\bP}{\bold{P}} \newcommand{\bQ}{\bold{Q}} \newcommand{\bR}{\bold{R}} \newcommand{\bS}{\bold{S}} \newcommand{\bT}{\bold{T}} \newcommand{\bU}{\bold{U}} \newcommand{\bV}{\bold{V}} \newcommand{\bW}{\bold{W}} \newcommand{\bX}{\bold{X}} \newcommand{\oX}{\overline{X}} \newcommand{\bY}{\bold{Y}} \newcommand{\bZ}{\bold{Z}} \newcommand{\calS}{\mathcal{S}} %%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%% % definitions for hyperref % %%%%%%%%%%%%%%%%%%%%%%%%%%%% %fix counters %\usepackage{aliascnt} %%%%%%%%%%%%%%%%%%%%%%%%%%%% % Optimization problems % %%%%%%%%%%%%%%%%%%%%%%%%%%%% \newcommand{\probmtflprimal}{\mathcal{E}_{\mbox{\tiny MTFL}}} \newcommand{\probmtfldual}{\mathcal{C}_{\mbox{\tiny MTFL}}} \newcommand{\probmtfldualeps}{\mathcal{C}_{\mbox{$\varepsilon$-\tiny {MTFL}}}} \newcommand{\mtflmodprimal}{\mathcal{E}} \newcommand{\mtflmoddual}{\mathcal{C}} \newcommand{\mtflmoddualeps}{{\mathcal{C}_{\vareps}}} \newcommand{\mtflmoddualmerge}{{\mathcal{S}_{\vareps}}} %%%%%%%%%%%%%%%%%%%%%%%%%%%% % Algorithms stuff % %%%%%%%%%%%%%%%%%%%%%%%%%%%% %\algnewcommand\algorithmicinput{\textbf{input:}} %\algnewcommand\Input{\item[\algorithmicinput]} %\algnewcommand\algorithmicoutput{\textbf{output:}} %\algnewcommand\Output{\item[\algorithmicoutput]} %\algdef{SE}[DOWHILE]{Do}{DoWhile}{\algorithmicdo}[1]{\algorithmicwhile\ #1} %%%%%%%%%%%%%%%%%%%%%%%%%%%% % Matrix operators % %%%%%%%%%%%%%%%%%%%%%%%%%%%% \newcommand{\transp}{\mathsf{T}} \DeclareMathOperator*{\argmin}{arg\,min} \DeclareMathOperator*{\argmax}{arg\,max} \DeclareMathOperator*{\trace}{trace} \DeclareMathOperator*{\range}{Ran} \DeclareMathOperator*{\kernel}{Ker} \DeclareMathOperator*{\diag}{Diag} \DeclareMathOperator*{\Vecm}{Vec} \DeclareMathOperator*{\Rank}{Rank} \DeclareMathOperator*{\Row}{Row} \DeclareMathOperator*{\Col}{Col} \DeclareMathOperator*{\Op}{\textbf{op}} \DeclareMathOperator*{\nnz}{\textbf{nnz}} %%%%%%%%%%%%%%%%%%%%%%%%%%%% % Norms % %%%%%%%%%%%%%%%%%%%%%%%%%%%% \newcommand{\norm}[2]{\left\Vert #1 \right\Vert_{#2}} \newcommand{\normsmall}[2]{\Vert #1 \Vert_{#2}} %%%%%%%%%%%%%%%%%%%%%%%%%%%% % Statistic operators % %%%%%%%%%%%%%%%%%%%%%%%%%%%% \newcommand{\probability}{\mathbb{P}} \newcommand{\probdist}{Pr} \DeclareMathOperator*{\expectedvalue}{\mathbb{E}} \newcommand{\expectedvalueover}[1]{\expectedvalue\limits_{#1}} \newcommand{\gaussdistr}{\mathcal{N}} \newcommand{\uniformdistr}{\mathcal{U}} \newcommand{\indvec}{J} %%%%%%%%%%%%%%%%%%%%%%%%%%%% % Algebraic Sets % %%%%%%%%%%%%%%%%%%%%%%%%%%%% \newcommand{\realnumbers}{\mathbb{R}} \newcommand{\realnumberssdp}{\mathbb{R}_{+}} \newcommand{\realnumbersdp}{\mathbb{R}_{++}} \newcommand{\naturalnumbers}{\mathbb{N}} \newcommand{\symmetricsdp}{\boldsymbol{S}^{d}_{+}} \newcommand{\symmetricdp}{\boldsymbol{S}^{d}_{++}} \newcommand{\symmetric}{\boldsymbol{S}^{d}} \newcommand{\orthogonal}{\boldsymbol{O}^{d}} %%%%%%%%%%%%%%%%%%%%%%%%%%%% % MDP related notation % %%%%%%%%%%%%%%%%%%%%%%%%%%%% \newcommand{\statespace}{\mathcal{X}} \newcommand{\statespacedisc}{S} \newcommand{\actionspace}{\mathcal{A}} \newcommand{\transitionspace}{\mathcal{P}} \newcommand{\transitionkernel}{P} \newcommand{\rewardspace}{\mathcal{R}} \newcommand{\rewardkernel}{R} \newcommand{\epochset}{\mathcal{T}} \newcommand{\trajspace}{\mathcal{H}} \newcommand{\funcspace}{\mathcal{F}} \newcommand{\dataset}{\mathcal{D}} \newcommand{\Qfuncspace}{\mathcal{Q}} \newcommand{\borelsetspace}{\mathcal{B}} \newcommand{\kernelspace}{\mathcal{K}} %\newcommand{\muset}{\boldsymbol\mu} \newcommand{\muset}{\mu_{\pi}} %\newcommand{\rhoset}{\boldsymbol\rho} \newcommand{\rhoset}{\rho_{\pi}} \newcommand{\rewardaverage}{R} \newcommand{\rewardsample}{r} \newcommand{\rewardsamplevec}{\br} %\newcommand{\optbellop}{\mathcal{T}^{*}} \newcommand{\optbellop}{\mathcal{T}} \newcommand{\T}{\mathcal{T}} \newcommand{\D}{\mathcal{P}} \renewcommand{\Re}{\mathbb{R}} %%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%% %Short version of notation % %%%%%%%%%%%%%%%%%%%%%%%%%%%% \newcommand{\wrt}{w.r.t. } %\newcommand{\defeq}{\stackrel{\mathclap{\normalfont\mbox{\tiny def}}}{=}} \newcommand{\maxund}[1]{\max\limits_{#1}} \newcommand{\minund}[1]{\min\limits_{#1}} \newcommand{\vareps}{\varepsilon} \newcommand{\bigotime}{\mathcal{O}} %%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%% %\newtheorem{theorem}{Theorem} %\newtheorem{lemma}{Lemma} %\newtheorem{corollary}{Corollary} %\newtheorem{proposition}{Proposition} %%\newtheorem{definition}{Definition} %\newtheorem{assumption}{Assumption} %%\newtheorem{remark}{Remark} \newcommand{\gl}{{\sc GL}\xspace} \newcommand{\mtfl}{{\sc MTFL}\xspace} \newcommand{\fqi}{{\sc F$Q$I}\xspace} \newcommand{\lassofqi}{{\sc LASSO--F$Q$I}\xspace} \newcommand{\glfqi}{{\sc GL--F$Q$I}\xspace} \newcommand{\flfqi}{{\sc FL--F$Q$I}\xspace} \newcommand{\lasso}{{\sc LASSO}\xspace} %\newaliascnt{lemma}{theorem} %\newtheorem{lemma}[lemma]{Lemma} %\aliascntresetthe{lemma} %\providecommand*{\lemmaautorefname}{Lemma} %\newaliascnt{corollary}{theorem} %\newtheorem{corollary}[corollary]{Corollary} %\aliascntresetthe{corollary} %\providecommand*{\corollaryautorefname}{Corollary} %\newaliascnt{proposition}{theorem} %\newtheorem{proposition}[proposition]{Proposition} %\aliascntresetthe{proposition} %\providecommand*{\propositionautorefname}{Proposition} %\newaliascnt{definition}{theorem} %\newtheorem{definition}[definition]{Definition} %\aliascntresetthe{definition} %\providecommand*{\definitionautorefname}{Definition} %\newaliascnt{assumption}{theorem} %\newtheorem{assumption}[assumption]{Assumption} %\aliascntresetthe{assumption} %\providecommand*{\assumptionautorefname}{Assumption} %\newenvironment{remark}[1][Remark]{\begin{trivlist} %\item[\hskip \labelsep {\bfseries #1}]}{\end{trivlist}}