From c78a1288e493510687204149aee429472f40ad90 Mon Sep 17 00:00:00 2001 From: Quarto GHA Workflow Runner Date: Sun, 24 Nov 2024 02:28:46 +0000 Subject: [PATCH] Built site for gh-pages --- .nojekyll | 2 +- index.html | 16 ++++++++-------- ...hed-202407-susmann-adaptive-conformal.html | 16 ++++++++-------- ...shed-202407-susmann-adaptive-conformal.pdf | Bin 1246537 -> 1246537 bytes .../figure-pdf/fig-aci-1.pdf | Bin 133583 -> 133583 bytes .../figure-pdf/fig-agaci-1.pdf | Bin 72230 -> 72230 bytes .../figure-pdf/fig-case-study-example-1.pdf | Bin 26791 -> 26791 bytes .../figure-pdf/fig-case-study-metrics-1.pdf | Bin 11299 -> 11299 bytes .../figure-pdf/fig-dtaci-example-1.pdf | Bin 133158 -> 133158 bytes .../figure-pdf/fig-saocp-1.pdf | Bin 136042 -> 136042 bytes .../figure-pdf/fig-sf-ogd-1.pdf | Bin 135522 -> 135522 bytes .../fig-simulation-one-example-1.pdf | Bin 19292 -> 19292 bytes .../figure-pdf/fig-simulation-one-joint-1.pdf | Bin 14055 -> 14055 bytes .../fig-simulation-one-results-1.pdf | Bin 21537 -> 21537 bytes .../fig-simulation-one-widths-1.pdf | Bin 79910 -> 79910 bytes .../fig-simulation-two-example-1.pdf | Bin 131464 -> 131464 bytes .../figure-pdf/fig-simulation-two-joint-1.pdf | Bin 10938 -> 10938 bytes .../fig-simulation-two-results-1.pdf | Bin 28475 -> 28475 bytes .../simulation_one_plot_appendix-1.pdf | Bin 50054 -> 50054 bytes .../figure-pdf/unnamed-chunk-4-1.pdf | Bin 40086 -> 40086 bytes ...ting-697306ee647f3aecb60be57249203282.css} | 2 +- 21 files changed, 18 insertions(+), 18 deletions(-) rename published-202407-susmann-adaptive-conformal_files/libs/quarto-html/{quarto-syntax-highlighting-722b7ecd703fa194e825a0d619ad685e.css => quarto-syntax-highlighting-697306ee647f3aecb60be57249203282.css} (97%) diff --git a/.nojekyll b/.nojekyll index d651cc5..0512189 100644 --- a/.nojekyll +++ b/.nojekyll @@ -1 +1 @@ -05063326 \ No newline at end of file +16022e28 \ No newline at end of file diff --git a/index.html b/index.html index e1ab7d2..0ca12f6 100644 --- a/index.html +++ b/index.html @@ -2,7 +2,7 @@ - + @@ -89,7 +89,7 @@ - + @@ -283,7 +283,7 @@
Modified
-

November 17, 2024

+

November 24, 2024

@@ -567,7 +567,7 @@

3 Algorithms

For demonstration purposes we assume we have access to unbiased predictions \hat{\mu}_t = 0 for all t \in \llbracket T \rrbracket. Throughout we set the target empirical coverage to \alpha = 0.8.

3.1 Adaptive Conformal Inference (ACI)

-
+
\begin{algorithm} \caption{Adaptive Conformal Inference} \begin{algorithmic} \State \textbf{Input:} starting value $\theta_1$, learning rate $\gamma > 0$. \For{$t = 1, 2, \dots, T$} \State \textbf{Output:} prediction interval $\widehat{C}_t(\theta_t)$. \State Observe $y_t$. \State Evaluate $\mathrm{err}_t = \mathbb{I}[y_t \not\in \widehat{C}_t(\theta_t)]$. \State Update $\theta_{t+1} = \theta_t + \gamma (\mathrm{err}_t - (1 - \alpha))$. \EndFor \end{algorithmic} \end{algorithm}
@@ -619,7 +619,7 @@

3.2 Aggregated Adaptive Conformal Inference (AgACI)

-
+
\begin{algorithm} \caption{Aggregated Adaptive Conformal Inference} \begin{algorithmic} \State \textbf{Input:} candidate learning rates $(\gamma_k)_{1 \leq k \leq K }$, starting value $\theta_1$. \State Initialize lower and upper BOA algorithms $\mathcal{B}^\ell := \texttt{BOA}(\alpha \leftarrow (1 - \alpha) / 2)$ and $\mathcal{B}^u := \texttt{BOA}(\alpha \leftarrow (1 - (1 - \alpha)/2))$. \For{$k = 1, \dots, K$} \State Initialize ACI $\mathcal{A}_k = \texttt{ACI}(\alpha \leftarrow \alpha, \gamma \leftarrow \gamma_k, \theta_1 \leftarrow \theta_1)$. \EndFor \For{$t = 1, 2, \dots, T$} \For{$k = 1, \dots, K$} \State Retrieve candidate prediction interval $[\ell^k_{t}, u^k_{t}]$ from $\mathcal{A}_k$. \EndFor \State Compute aggregated lower bound $\tilde{\ell}_t := \mathcal{B}^\ell((\ell^k_t : k \in \{ 1, \dots, K \}))$. \State Compute aggregated upper bound $\tilde{u}_t := \mathcal{B}^u((u^k_t : k \in \{ 1, \dots, K \}))$. \State \textbf{Output:} prediction interval $[\tilde{\ell}_t, \tilde{u}_t]$. \State Observe $y_t$. \For{$k = 1, \dots, K$} \State Update $\mathcal{A}_k$ with observation $y_t$. \EndFor \State Update $\mathcal{B}^\ell$ with observed outcome $y_t$. \State Update $\mathcal{B}^u$ with observed outcome $y_t$. \EndFor \end{algorithmic} \end{algorithm}
@@ -650,7 +650,7 @@

3.3 Dynamically-tuned Adaptive Conformal Inference (DtACI)

-
+
\begin{algorithm} \caption{Dynamically-tuned Adaptive Conformal Inference} \begin{algorithmic} \State \textbf{Input:} starting value $\theta_1$, candidate learning rates $(\gamma_k)_{1 \leq k \leq K }$, parameters $\sigma, \eta$. \For{$k = 1, \dots, K$} \State Initialize expert $\mathcal{A}_k = \texttt{ACI}(\alpha \leftarrow \alpha, \gamma \leftarrow \gamma_k, \theta_1 \leftarrow \theta_1)$. \EndFor \For{$t = 1, 2, \dots, T$} \State Define $p_t^k := p_t^k / \sum_{i=1}^K p_t^i$, for all $1 \leq k \leq K$. \State Set $\theta_t = \sum_{k=1}^K \theta_t^k p_t^k$. \State \textbf{Output:} prediction interval $\widehat{C}_t(\theta_t)$. \State Observe $y_t$ and compute $r_t$. \State $\bar{w}_{t}^k \gets p_t^k \exp(-\eta L^\alpha(\theta_t^k, r_t))$, for all $1 \leq k \leq K$. \State $\bar{W}_t \gets \sum_{i=1}^K \bar{w}_t^i$. \State $p_{t+1}^k \gets (1 - \sigma) \bar{w}_t^k + \bar{W}_t \sigma / K$. \State Set $\mathrm{err}_t := \mathbb{I}[y_t \not\in \widehat{C}_t(\theta_t)]$. \For{$k = 1, \dots, K$} \State Update ACI $\mathcal{A}_k$ with $y_t$ and obtain $\theta_{t+1}^k$. \EndFor \EndFor \end{algorithmic} \end{algorithm}
@@ -703,7 +703,7 @@

3.4 Scale-Free Online Gradient Descent (SF-OGD)

-
+
\begin{algorithm} \caption{Scale-Free Online Gradient Descent} \begin{algorithmic} \State \textbf{Input:} starting value $\theta_1$, learning rate $\gamma > 0$. \For{$t = 1, 2, \dots, T$} \State \textbf{Output:} prediction interval $\widehat{C}_t(\theta_t)$. \State Observe $y_t$ and compute $r_t$. \State Update $\theta_{t+1} = \theta_t - \gamma \frac{\nabla L^\alpha(\theta_t, r_t)}{\sqrt{\sum_{i=1}^t} \| \nabla L^\alpha(\theta_i, r_i) \|_2^2}$. \EndFor \end{algorithmic} \end{algorithm}
@@ -742,7 +742,7 @@

3.5 Strongly Adaptive Online Conformal Prediction (SAOCP)

-
+
\begin{algorithm} \caption{Strongly Adaptive Online Conformal Prediction} \begin{algorithmic} \State \textbf{Input:} initial value $\theta_0$, learning rate $\gamma > 0$. \For{$t = 1, 2, \dots, T$} \State Initialize expert $\mathcal{A}_t = \texttt{SF-OGD}(\alpha \leftarrow \alpha, \gamma \leftarrow \gamma, \theta_1 \leftarrow \theta_{t-1})$, set weight $p_t^t = 0$. \State Compute active set $\mathrm{Active}(t) = \{ i \in \llbracket T \rrbracket : t - L(i) < i \leq t \}$ (see below for definition of $L(t)$). \State Compute prior probability $\pi_i \propto i^{-2} (1 + \lfloor \log_2 i \rfloor )^{-1} \mathbb{I}[i \in \mathrm{Active}(t)]$. \State Compute un-normalized probability $\hat{p}_i = \pi_i [p_{t,i}]_+$ for all $i \in \llbracket t \rrbracket$. \State Normalize $p = \hat{p} / \| \hat{p} \|_1 \in \Delta^t$ if $\| \hat{p} \|_1 > 0$, else $p = \pi$. \State Set $\theta_t = \sum_{i \in \mathrm{Active}(t)} p_i \theta_t^i$ (for $t \geq 2$), and $\theta_t = 0$ for $t = 1$. \State \textbf{Output:} prediction set $\widehat{C}_t(\theta_t)$. \State Observe $y_t$ and compute $r_t$. \For{$i \in \mathrm{Active}(t)$} \State Update expert $\mathcal{A}_t$ with $y_t$ and obtain $\theta_{t+1}^i$. \State Compute $g_t^i = \begin{cases} \frac{1}{D}\left(L^\alpha(\theta_t, r_t) - L^\alpha(\theta_t^i, r_t)\right) & p_t^i > 0 \\ \frac{1}{D}\left[L^\alpha(\theta_t, r_t) - L^\alpha(\theta_t^i, r_t))\right]_+ & p_t^i \leq 0 \\ \end{cases}$. \State Update expert weight $p_{t+1}^i = \frac{1}{t - i + 1}\left( \sum_{j=i}^t g_j^i \right) \left(1 + \sum_{j=i}^t p_j^i g_j^i \right)$. \EndFor \EndFor \end{algorithmic} \end{algorithm}
diff --git a/published-202407-susmann-adaptive-conformal.html b/published-202407-susmann-adaptive-conformal.html index e1ab7d2..0ca12f6 100644 --- a/published-202407-susmann-adaptive-conformal.html +++ b/published-202407-susmann-adaptive-conformal.html @@ -2,7 +2,7 @@ - + @@ -89,7 +89,7 @@ - + @@ -283,7 +283,7 @@
Modified
-

November 17, 2024

+

November 24, 2024

@@ -567,7 +567,7 @@

3 Algorithms

For demonstration purposes we assume we have access to unbiased predictions \hat{\mu}_t = 0 for all t \in \llbracket T \rrbracket. Throughout we set the target empirical coverage to \alpha = 0.8.

3.1 Adaptive Conformal Inference (ACI)

-
+
\begin{algorithm} \caption{Adaptive Conformal Inference} \begin{algorithmic} \State \textbf{Input:} starting value $\theta_1$, learning rate $\gamma > 0$. \For{$t = 1, 2, \dots, T$} \State \textbf{Output:} prediction interval $\widehat{C}_t(\theta_t)$. \State Observe $y_t$. \State Evaluate $\mathrm{err}_t = \mathbb{I}[y_t \not\in \widehat{C}_t(\theta_t)]$. \State Update $\theta_{t+1} = \theta_t + \gamma (\mathrm{err}_t - (1 - \alpha))$. \EndFor \end{algorithmic} \end{algorithm}
@@ -619,7 +619,7 @@

3.2 Aggregated Adaptive Conformal Inference (AgACI)

-
+
\begin{algorithm} \caption{Aggregated Adaptive Conformal Inference} \begin{algorithmic} \State \textbf{Input:} candidate learning rates $(\gamma_k)_{1 \leq k \leq K }$, starting value $\theta_1$. \State Initialize lower and upper BOA algorithms $\mathcal{B}^\ell := \texttt{BOA}(\alpha \leftarrow (1 - \alpha) / 2)$ and $\mathcal{B}^u := \texttt{BOA}(\alpha \leftarrow (1 - (1 - \alpha)/2))$. \For{$k = 1, \dots, K$} \State Initialize ACI $\mathcal{A}_k = \texttt{ACI}(\alpha \leftarrow \alpha, \gamma \leftarrow \gamma_k, \theta_1 \leftarrow \theta_1)$. \EndFor \For{$t = 1, 2, \dots, T$} \For{$k = 1, \dots, K$} \State Retrieve candidate prediction interval $[\ell^k_{t}, u^k_{t}]$ from $\mathcal{A}_k$. \EndFor \State Compute aggregated lower bound $\tilde{\ell}_t := \mathcal{B}^\ell((\ell^k_t : k \in \{ 1, \dots, K \}))$. \State Compute aggregated upper bound $\tilde{u}_t := \mathcal{B}^u((u^k_t : k \in \{ 1, \dots, K \}))$. \State \textbf{Output:} prediction interval $[\tilde{\ell}_t, \tilde{u}_t]$. \State Observe $y_t$. \For{$k = 1, \dots, K$} \State Update $\mathcal{A}_k$ with observation $y_t$. \EndFor \State Update $\mathcal{B}^\ell$ with observed outcome $y_t$. \State Update $\mathcal{B}^u$ with observed outcome $y_t$. \EndFor \end{algorithmic} \end{algorithm}
@@ -650,7 +650,7 @@

3.3 Dynamically-tuned Adaptive Conformal Inference (DtACI)

-
+
\begin{algorithm} \caption{Dynamically-tuned Adaptive Conformal Inference} \begin{algorithmic} \State \textbf{Input:} starting value $\theta_1$, candidate learning rates $(\gamma_k)_{1 \leq k \leq K }$, parameters $\sigma, \eta$. \For{$k = 1, \dots, K$} \State Initialize expert $\mathcal{A}_k = \texttt{ACI}(\alpha \leftarrow \alpha, \gamma \leftarrow \gamma_k, \theta_1 \leftarrow \theta_1)$. \EndFor \For{$t = 1, 2, \dots, T$} \State Define $p_t^k := p_t^k / \sum_{i=1}^K p_t^i$, for all $1 \leq k \leq K$. \State Set $\theta_t = \sum_{k=1}^K \theta_t^k p_t^k$. \State \textbf{Output:} prediction interval $\widehat{C}_t(\theta_t)$. \State Observe $y_t$ and compute $r_t$. \State $\bar{w}_{t}^k \gets p_t^k \exp(-\eta L^\alpha(\theta_t^k, r_t))$, for all $1 \leq k \leq K$. \State $\bar{W}_t \gets \sum_{i=1}^K \bar{w}_t^i$. \State $p_{t+1}^k \gets (1 - \sigma) \bar{w}_t^k + \bar{W}_t \sigma / K$. \State Set $\mathrm{err}_t := \mathbb{I}[y_t \not\in \widehat{C}_t(\theta_t)]$. \For{$k = 1, \dots, K$} \State Update ACI $\mathcal{A}_k$ with $y_t$ and obtain $\theta_{t+1}^k$. \EndFor \EndFor \end{algorithmic} \end{algorithm}
@@ -703,7 +703,7 @@

3.4 Scale-Free Online Gradient Descent (SF-OGD)

-
+
\begin{algorithm} \caption{Scale-Free Online Gradient Descent} \begin{algorithmic} \State \textbf{Input:} starting value $\theta_1$, learning rate $\gamma > 0$. \For{$t = 1, 2, \dots, T$} \State \textbf{Output:} prediction interval $\widehat{C}_t(\theta_t)$. \State Observe $y_t$ and compute $r_t$. \State Update $\theta_{t+1} = \theta_t - \gamma \frac{\nabla L^\alpha(\theta_t, r_t)}{\sqrt{\sum_{i=1}^t} \| \nabla L^\alpha(\theta_i, r_i) \|_2^2}$. \EndFor \end{algorithmic} \end{algorithm}
@@ -742,7 +742,7 @@

3.5 Strongly Adaptive Online Conformal Prediction (SAOCP)

-
+
\begin{algorithm} \caption{Strongly Adaptive Online Conformal Prediction} \begin{algorithmic} \State \textbf{Input:} initial value $\theta_0$, learning rate $\gamma > 0$. \For{$t = 1, 2, \dots, T$} \State Initialize expert $\mathcal{A}_t = \texttt{SF-OGD}(\alpha \leftarrow \alpha, \gamma \leftarrow \gamma, \theta_1 \leftarrow \theta_{t-1})$, set weight $p_t^t = 0$. \State Compute active set $\mathrm{Active}(t) = \{ i \in \llbracket T \rrbracket : t - L(i) < i \leq t \}$ (see below for definition of $L(t)$). \State Compute prior probability $\pi_i \propto i^{-2} (1 + \lfloor \log_2 i \rfloor )^{-1} \mathbb{I}[i \in \mathrm{Active}(t)]$. \State Compute un-normalized probability $\hat{p}_i = \pi_i [p_{t,i}]_+$ for all $i \in \llbracket t \rrbracket$. \State Normalize $p = \hat{p} / \| \hat{p} \|_1 \in \Delta^t$ if $\| \hat{p} \|_1 > 0$, else $p = \pi$. \State Set $\theta_t = \sum_{i \in \mathrm{Active}(t)} p_i \theta_t^i$ (for $t \geq 2$), and $\theta_t = 0$ for $t = 1$. \State \textbf{Output:} prediction set $\widehat{C}_t(\theta_t)$. \State Observe $y_t$ and compute $r_t$. \For{$i \in \mathrm{Active}(t)$} \State Update expert $\mathcal{A}_t$ with $y_t$ and obtain $\theta_{t+1}^i$. \State Compute $g_t^i = \begin{cases} \frac{1}{D}\left(L^\alpha(\theta_t, r_t) - L^\alpha(\theta_t^i, r_t)\right) & p_t^i > 0 \\ \frac{1}{D}\left[L^\alpha(\theta_t, r_t) - L^\alpha(\theta_t^i, r_t))\right]_+ & p_t^i \leq 0 \\ \end{cases}$. \State Update expert weight $p_{t+1}^i = \frac{1}{t - i + 1}\left( \sum_{j=i}^t g_j^i \right) \left(1 + \sum_{j=i}^t p_j^i g_j^i \right)$. \EndFor \EndFor \end{algorithmic} \end{algorithm}
diff --git a/published-202407-susmann-adaptive-conformal.pdf b/published-202407-susmann-adaptive-conformal.pdf index e5cb722320bcb45a133cc24cda1a2bad532ff601..b8d33268e0865f7a0cded7c34a16c00f577e4885 100644 GIT binary patch delta 3335 zcma)4dpy(c9%p8l%aB_rVq_y`!%Q^7TqZJ#ZZ1i`$YrEl!e%O&a!V$JWYvkZ6#JFM z79k~_$SrceQ@NF49Hq`V|LC08=kg0J8gvXuD)E__0<+X``*t{x#ByE~QWo4LE>0vsjp zc<+14(Gh63Qe@pAA?x^r@T96rX7}h5Wr4S69O#D#vP###4cezyb5G6_CaA&B0%Oi! zz)-cVAHN8C`Mfv!WvP9J1J=xa7ok$UDGz=V=`dc}Kr$_ZPw$GS9t+;BH`)D0gVe&C zS-7%j7T(EI!N0bC0GK~gPCYl)cZo0NDgJXbWB06f{fyCc=560Y|Jad`JM+vwON^>* z2zq3CSJla9Q97$+*@YY`_3+SqMENYX6S7P0YG`iu@w-*o=ff|v!re*7k7dG-vO?@` z?Qm;cnKU8grY&o>igvw1`W$^!R;TK-NFHi{Y0CDFy6yP%M#6_;nR2{*b#bA3MFVoN zuF|pYe&N=nMQkEVad&*`>%p6G0obC&wL)0ZSI>4CH|l*_ z_CSJL2++}~lIiNZGG%Ntpga@fsDEN^$I#jxq|Pi!{2@F(vC-=9{UEJIFY^mV(SZR^ zPnsq13QOs|*Vr#1%1bpLe9|I1p3{g+a1W0=4R$ueXJfG3`uwk)-mi1KgAciZ)orDE z3di_yJ>=@@K+^MNWNDw5!*=x|SObS$kIOA*0k+d<21~`;Fs+uIFy#BrAWW+QDbZq* zr{E~{h)V~A%84`=l@A^eiL4vOm4@(|GG?=FiK3X;s2dM=3RLx)aC+P3b5{&#c?*Z~ zAk)<0iAwF2Y^~OICCjG~3J+Oa$_pMu!>Fe6v6{GfJQQhwKqAmcW0}V1kylX|2IbKQxkQj5m^9}c&W4msWg73R2pCc4%)zJ{49W+@Wvj6_{mLJ=Q4>b z0fW8jzs`*|&yS+@s)-VQrcZEL6IfqBrpuaO>egSiq0oQOqW@hRCLwJrXnhGW)3ypz zG=7TpQ3%9l;vd+5i8)x|1P)e&$-#nK6n<8{PSF2@n6}mjNn-9j1de+3`!3WU=qTp8 ziBR|7RD(ZI4K_TCJTJxCq=B)}bxX*9Vrsl1;w_zUBTv2?6_?y4! zAguF^H~ji>NkVUu`1+Dy_9j8-8!v@KBRB2p2lhiuU@p-i0~;N!pBZlE!Vj#XA4046 zS=Ra=<_}EFMrQFWNkPufnM9)hGm}U(3X$0lp)%Qrq$W|n(xEne*mb)8hW+bZGF6kM z;YT!nj4;yJ0C9NJ8rDH(@VK;_3=V@p=^LW;EpY}GXpDvRUQ2`}PTvB7#95;(&{ime zJ*uYvn{td6Vk*2HsE;;8r&-95G&228Z95=SKgyyNjlVv!6Q83Q*mh@MPQ@L+77aWN zdMEI`q2|68BfkgkwSuzNu2v!Ze#uJ~NB_?EKVT&(MOX2!cGuP_EEoX8DQaMfk}|NJ zyRCX93-#B@PdAs>dX_b*g=4+0&^EEd=BMG*Ms+*0PQ?(Y^l%hUOT5vXB#{vhX7IUE zE~*OLHmZyA$* z$LC1Cw$cS+NX=9Bd7(N0OPr|0uHf(AI~vF2$p-HAPq(v$8-<`=;C+*i!yqPMY=qlENl5p^lx0!5wF z*5dM5iFQim$Sno{psl+8xgm7Wyuog1y0Djn6fPm}S09JOq{b~sDMEQpDYTM{QqMFamrdlC{MM?^}cav@N5Nb0*d*!drMA@0$?3 zJ?^?Zh>s(5MYx90N$RRBp|btOjAd{f-B)5UBze1&&zkf&@d(`;G3ZDPKVqksop})% z(@#E8wbr9&kwk`j+as$#4UW2p@h1SbUl=4@UnR{yt{ros!24zJS|PX3>I_^CgyyE+ zpF(?phwYNf3Jy}$^P_yWHpD$~HFV-Wh3^M7S_r!H!8vqQ@$_>1)j`~+V*OQzlQ!gR1zE>+Kyhu*UKd#-ww4-H)P_(s-cfJy`> z*|G&nCA99Fqbmb}1#Y2`bCI18qqruLmuLrgsfgGt8vDu-Ay$F!?+?PFf#=B1FjLRz z(R1-qFw>ZZ*T-s`#^&58B1DCwg5PZWyX#|nb|64?91jL$N=S%p0b^lg?D}&idCBTt z1n`&VUxp2`0H0qgeeT>uMHZ8AmLU!sKiGR6fGNo@>P>kRD3i=ug^pci!!#{G8J>yX3DgO<;*_jFy$+ zlupBw?QQ#!)3}C_{3BN0TE0CQwLZ0hW*I{#t2_gXp!gcYk5+b7mxvV?)8B|sl1fsf zAz}*wWNEaRNT8BNmBRQ(343Ma1=Jef3vEkdrCYUzXRZ*`ZgIC&!gwa66$OpD>=a); zI<{4)Dqp(3sdrYDp2EI+f|9=Ay@zxyW6#n2shSMofO~)0Xz_YdKO_^%TE%ux8`GVD z=em?N@b5bwPC@C-iYGt98f^zuNLZ?|bZYwhNfoHH$CMLs+&efUE;KbFio04A!c73G ztPSgc{Lc$Pazo}G_Sl`vLF2Ttp`AI0Sx11Y)^hmnwp@d*ku6ik zQ$33NMJJ%!tM>Sbk@ET!XFea_56QHRp02bT40EF7tj+|f6NhPb(Rh-R16L9OQpuCpF7{BB&mv4APZhg~xV$-yL6<2ChHm$S#P@)F#)G!?-uYQjcf K4Gn829Qa>9W|7JO delta 3104 zcmai#dpy(a8^?`p<`^?NkQrOW&zYI2Fw9ww%OQtR$RSG4Bj*j3wjrmGs=R+io$zkGIdV1>jN59|eeqDcDuj_q(?$3Q)_v`y@q1v`kZ8MDcG1%KU zeXsy~uN@;E>;)t(b9xePnwpQX)IGJ7@b0?$)GoJ{y*W43?hltQP~UM&u2w5KoP{rjH(bqh|5^j+3c4cJ%iPn%iNn z!J>ZA2XIBtKA|+-m#*=RD|$1g*umqZJ8%!x9pn1w#sl%_rH>NPbM*D64^#;?QKMNV zrnhoddVkj^DJ|KRZ@ruC|E$K(KMnWn6QCv2*;eIoN0kZd;?8Hl6#2&BFYFo8ijBN* z+NsN_1iX$6yCE#FOL}hH@nxs9#e*zoSO!XjZr|NETxlTCmte`>z&pzf*WOaT1tVyn z8@E)AKnjZ@!z)^s`%a`k#g1?b{hB9Cn>XzI-Ns}|-UP7J18&%X3G?TThLdZDC%Kj03;hX|UOt)+ z_){MMNTmMtm?eMD`uhBR=9bPG$%*IF^h2US`_wgpIIepm%2Vz*A1P(LjVRYzhbH%M zwG{3A7HmLcs?Z{`IXO{u#zT z_#5fxuTHrxqaDkuFlQs!n%>b1pfg-}2&~}|11V;l4AH`9YHR9g>*rm`lYsrm5G;OL zvQ%b@R3?BamB~M?@=r-ea}xxA+m-@n@iXBpffSSPWxbzuboBDR>ZA}h1b6{t5;i~? z+rN{T{~|G8Noq>=b}%mlnc3SRNvKUyOIuUt$B|mvc|(dK&UvyM0@#!WBM2{qm<>k4 z-P?W*Gez8m87<%IKN;jnZyI!D>L`hHyZ#TtDJ3pegc%nr8g1|^>P4C)m3g&VRORqL zgx0rC{hKk_+r>;y{G0x`y_Pl){klD6m?p*xapo{>OWC(RY3u1~{UA3pUY{J9+_EjB z->m$C$m3la%S;{KB38Y5Bo9Mx3JHFl952fQ3~zB_isws!>%Rj(O@s$*PQ>w>&J%aZ z8J%BXDLm`U7irMgLqeO<7__d(PnLNp^Td0oE6iNqR8I?IYGS5^#_5@&@jAMAJzcCe z27@ufX`%6YI))1S{y*ZChl7eK@T0MrI2Z|)>wZ+803%PXkr`|d17tL(UY>=&LO9yw@0!UZ~ za4QZ}4s^zWT@qExLEXkX^Tns@MLYR+)Z2Cf3veRmP}JbH;b;d-d`0Y5GPZc9ZZdd? zsgVTr-C7BAkk@9`w9kxKr$45`l=G0@=M1pQEnpam;j zlrG_)tqLpx+soxnt>MFW3XBr@5NPV)y%{g#=K)0$so)t<)&KmHQfg1=qC^8Qts+ zX@*auz8t=^bQ?|L??)9ME*1vtHoZ~0k5K=cnZUoR8RF%1mZpJoQ8S52}lzhpt z;#A+FoNMZ_a2Y|4cnIGmBt7s+gz_JHO(_bNb%kUs;N%mPiqNPQMo3x2EGpcQMVI7Q zkB1uV#k@5^{fI$PmwZ@JoAW$$e;FY39apJc*}Q^ zKNn~woG%q=J&`gP5AguH2*sinxXJID{uI%NeNvFiz0TQSz^|n~MwUZ&^_fs)k6lvo zkW(5y=i*=c*#UCCt2N0(S=@WkOsYBYuFJg-?8hpapq#`A2Y^bwOGf7P(WPDXu|p+3 zT4G~m(_thw9Sk#;%rO?dj^f67A^Upmx%%m!;w~TYN$6>~NHi!HlmkVlC&gVovN5%Y zm@OvU*qOjdot1AFVAMqzZd>n)&)Una4s<1hvS@kHHKM9O zTAg2~P%c@W?h>)&GhaVr%p8+1nDfHLe+~$=72uoOxibAF^7X z0rTDyJs6<5=ghUpm>Noi4tGtBh6X3kCq&ASE zNrN!u>0&03nC+Jxz~`s3D_FkxykELsdPz=Cl7VsjJx}VWJ@?BeyevRmd68hKN^PYE zwEmGaLL145Q@NwEZWqR${u}@A4zm1DP6sO zR^!_;$5kTBy3cc^_Ym(n8DP2_E`}m{I`&i|vK&23;iENL>9Qd?T{O%W;SHawN6u>K z(K>ykef}c7HxC-7wJDX}J$reA6o2#9?FZH8GW;>r(DjFSWzn*mw2b*}wqN!&rDEwrXXSFih4+d<7q&<&22nT@fvb$CLd z<#JB}aBklRigK|ivd9~I8F_0hV`#dgsFL@2R-9S4@ZG##cYhb45~~+wYE= z6z^yUg>mC3RRcU*p7x*m&>iYP8giZ0($EMc5k@P+hCZePG)4hK())m{sP)4-2=-Eo d;xkM-egmL(dvz=i4w6!Y=nA7yczbi=bcMka=a5V^+4)=0*!kxXTr E0Hrhzz5oCK delta 53 zcmX>=bcMka=a5V^+4)=0*!kxXTr E0Ht&ezyJUM diff --git a/published-202407-susmann-adaptive-conformal_files/figure-pdf/fig-agaci-1.pdf b/published-202407-susmann-adaptive-conformal_files/figure-pdf/fig-agaci-1.pdf index cb03fbcd9a7ee2ae7ad2bdc1162f9a70d30ec0ed..05bbec17d8a17cc0bf10327cd1c1db00c596e966 100644 GIT binary patch delta 50 zcmZ3sg=N_mmI>xcMkWSE2Ij^Fnq2z6`6(`mC8-J;E>=bcMka=a5V^+4t&xmpB>{A| B4kiEq delta 50 zcmZ3sg=N_mmI>xchUNxF24==bcMka=a5V^+4t&xmpB>{Bt B4k!Qs diff --git a/published-202407-susmann-adaptive-conformal_files/figure-pdf/fig-case-study-example-1.pdf b/published-202407-susmann-adaptive-conformal_files/figure-pdf/fig-case-study-example-1.pdf index 0566b6e40e62e105860d656d8d4f7bce2651f125..c5ae38730fe6180bf2b12338d8788a637e7a30ad 100644 GIT binary patch delta 49 ycmZ2}k#YG&#tG&sMkWSEM&?GQnq2z6`6(`mC8-J;E>=bcMka=aFu9G=bcMka=aFu9GN!zHfetOJYf?f`*Hgk%5tkp&?9eWAq7a06!WG{{R30 diff --git a/published-202407-susmann-adaptive-conformal_files/figure-pdf/fig-dtaci-example-1.pdf b/published-202407-susmann-adaptive-conformal_files/figure-pdf/fig-dtaci-example-1.pdf index 41b00c304e7e374a67551b7e4771558303c04323..7fe537218f7d18f3eac582409b85b011f4aa0ba3 100644 GIT binary patch delta 53 zcmZ2Bfn(VOjtS;UMkWSE2Ij_Qnq2z6`6(`mC8-J;E>=bcMka=a5V^+4)=0*!kxXaV E0hA67x&QzG delta 53 zcmZ2Bfn(VOjtS;UhUNxF24==)nq2z6`6(`mC8-J;E>=bcMka=a5V^+4)=0*!kxXaV E0hCS-yZ`_I diff --git a/published-202407-susmann-adaptive-conformal_files/figure-pdf/fig-saocp-1.pdf b/published-202407-susmann-adaptive-conformal_files/figure-pdf/fig-saocp-1.pdf index 6f3ccbe54b01b7c72350c1a3e73b080b46b3bc0d..3ff771509f7baef3849899328c4c4d45f2b5c29b 100644 GIT binary patch delta 55 zcmaF0jN{cZjtS-}MkWSEhQ=bcMka=aFuBI))@a79(M=bcMka=aFuBI))@a79(MBjtS;UMkWSE2Ij`*nq2z6`6(`mC8-J;E>=bcMka=a5V^+4)=0*!kxZO| E0I~B9od5s; delta 53 zcmaE~h~v>BjtS;UhUNxF24=?Qnq2z6`6(`mC8-J;E>=bcMka=a5V^+4)=0*!kxZO| E0J1X=bcMka=aFu9G=bcMka=aFu9G=bcMka=aFu9G<$ASS}ehsq# delta 49 ycmZ3uf^p#r#tG&shUNxFhDHWvnq2z6`6(`mC8-J;E>=bcMka=aFu9G<$ASS~C=I&+ diff --git a/published-202407-susmann-adaptive-conformal_files/figure-pdf/fig-simulation-one-widths-1.pdf b/published-202407-susmann-adaptive-conformal_files/figure-pdf/fig-simulation-one-widths-1.pdf index a2ceaa7e17801d697648f5d2ee6c247a1e671773..bffbee12f77d6680e4a496f23f14e3e1780d7275 100644 GIT binary patch delta 52 zcmZ4Xfo0hTmI>x6MkWSEhQ=bcMka=aFuBI)tkwY B4<7&k delta 52 zcmZ4Xfo0hTmI>x6hUNxFhDHWvnq2z6`6(`mC8-J;E>=bcMka=aFuBI)tkyE B42OlB%HLVr67tWMXItlWUA_jb_{$&7{i+ E0Go#nCjbBd delta 55 zcmeBZ=IChVm|(7AXl`I+Xk=)h$))d`pW>2OlB%HLVr67tWMXItlWUA_jb_{$&7{i+ E0Ge+O8vpoxr&jAfsv89v7siHzHfetOJYf?f`*Hgk%5tkp&?9eV|2V007V@Pxc~qF delta 47 wcmdlLx+`>oxr(8=fsv7!iMb}1zHfetOJYf?f`*Hgk%5tkp&?9eV|2V007hX9$p8QV diff --git a/published-202407-susmann-adaptive-conformal_files/figure-pdf/fig-simulation-two-results-1.pdf b/published-202407-susmann-adaptive-conformal_files/figure-pdf/fig-simulation-two-results-1.pdf index 93624bb3fe4dde6cfde13ccf8bdefc421a6191f2..98c3ac0b1c00d7212f63d956e8ab74bd440f352b 100644 GIT binary patch delta 49 ycmdmek8$@s#tG&sMkWSEhQ=bcMka=aFu9G<5ApzesSc0; delta 49 ycmdmek8$@s#tG&shUNxFhDL@4nq2z6`6(`mC8-J;E>=bcMka=aFu9G<5Apzd{|<%# diff --git a/published-202407-susmann-adaptive-conformal_files/figure-pdf/simulation_one_plot_appendix-1.pdf b/published-202407-susmann-adaptive-conformal_files/figure-pdf/simulation_one_plot_appendix-1.pdf index 949517082f9a6fa3292e1a3a56835ea102d93676..9bac19b6c18004cb0458a2556d538d4f5f010843 100644 GIT binary patch delta 49 ycmZo`XKrg}o?xzGWMW`sWNu`s$))d`pW>2OlB%HLVr67tWMXItliL`reHZ{qbPX;5 delta 49 ycmZo`XKrg}o?xzGXl`I+WM*Qj$))d`pW>2OlB%HLVr67tWMXItliL`reHZ{qWDP6; diff --git a/published-202407-susmann-adaptive-conformal_files/figure-pdf/unnamed-chunk-4-1.pdf b/published-202407-susmann-adaptive-conformal_files/figure-pdf/unnamed-chunk-4-1.pdf index e81fdf14e0c94e52a9025cf4bd1229cc93923f1a..4150b734a8b69df576b3b50108e5dd6d627fd889 100644 GIT binary patch delta 49 ycmbQXlWE#crU~XMMkWSEhQ=bcMka=aFu9G<)^h+}*bRFC delta 49 ycmbQXlWE#crU~XMhUNxFhDHX)nq2z6`6(`mC8-J;E>=bcMka=aFu9G<)^h+~Tn&K$ diff --git a/published-202407-susmann-adaptive-conformal_files/libs/quarto-html/quarto-syntax-highlighting-722b7ecd703fa194e825a0d619ad685e.css b/published-202407-susmann-adaptive-conformal_files/libs/quarto-html/quarto-syntax-highlighting-697306ee647f3aecb60be57249203282.css similarity index 97% rename from published-202407-susmann-adaptive-conformal_files/libs/quarto-html/quarto-syntax-highlighting-722b7ecd703fa194e825a0d619ad685e.css rename to published-202407-susmann-adaptive-conformal_files/libs/quarto-html/quarto-syntax-highlighting-697306ee647f3aecb60be57249203282.css index 04af414..c013f19 100644 --- a/published-202407-susmann-adaptive-conformal_files/libs/quarto-html/quarto-syntax-highlighting-722b7ecd703fa194e825a0d619ad685e.css +++ b/published-202407-susmann-adaptive-conformal_files/libs/quarto-html/quarto-syntax-highlighting-697306ee647f3aecb60be57249203282.css @@ -176,4 +176,4 @@ code span.wa { content: "