Commit 4453a8a3 by 周航

Fine-tune the figures

parent b18fa224
...@@ -47,32 +47,37 @@ brace/.style={ ...@@ -47,32 +47,37 @@ brace/.style={
% trajectory nodes % trajectory nodes
\node[box] (plan) { \node[box] (plan) {
\textbf{\textit{Plan}}\\ \colorbox{blue!20}{\textbf{\textit{Plan}}}\\
$p$ \vspace{.5ex}
$p$
}; };
\node[box, right=0.65cm of plan] (step1) { \node[box, right=0.65cm of plan] (step1) {
\textbf{\textit{Step 1}}\\ \colorbox{red!20}{\textbf{\textit{Step 1}}}\\
$u_1 \rightarrow o_1$ \vspace{.5ex}
$u_1 \rightarrow o_1$
}; };
\node[box, right=0.65cm of step1] (step2) { \node[box, right=0.65cm of step1] (step2) {
\textbf{\textit{Step 2}}\\ \colorbox{red!20}{\textbf{\textit{Step 2}}}\\
$u_2 \rightarrow o_2$ \vspace{.5ex}
$u_2 \rightarrow o_2$
}; };
\node[right=0.35cm of step2] (dots) { \node[right=0.35cm of step2] (dots) {
$\cdots$ $\cdots$
}; };
\node[box, right=0.35cm of dots] (stepT) { \node[box, right=0.35cm of dots] (stepT) {
\textbf{\textit{Step T}}\\ \colorbox{red!20}{\textbf{\textit{Step T}}}\\
$u_T \rightarrow o_T$ \vspace{.5ex}
$u_T \rightarrow o_T$
}; };
\node[rewardbox, right=0.75cm of stepT] (finalreward) { \node[rewardbox, right=0.75cm of stepT] (finalreward) {
\textbf{\textit{Final Reward}}\\ \colorbox{green!20}{\textbf{\textit{Final Reward}}}\\
$R(\tau)$ \vspace{.5ex}
$R(\tau)$
}; };
% arrows on trajectory % arrows on trajectory
...@@ -119,22 +124,22 @@ $r_T$ ...@@ -119,22 +124,22 @@ $r_T$
% curved credit-assignment arrows % curved credit-assignment arrows
\draw[dashedarrow] \draw[dashedarrow]
(assign.south west) (assign.south)
.. controls +(down:1.25cm) and +(down:1.25cm) .. .. controls +(down:1.25cm) and +(down:1.25cm) ..
(rT.south); (rT.south);
\draw[dashedarrow] \draw[dashedarrow]
(assign.south west) (assign.south)
.. controls +(down:1.55cm) and +(down:1.55cm) .. .. controls +(down:1.55cm) and +(down:1.55cm) ..
(r2.south); (r2.south);
\draw[dashedarrow] \draw[dashedarrow]
(assign.south west) (assign.south)
.. controls +(down:1.85cm) and +(down:1.85cm) .. .. controls +(down:1.85cm) and +(down:1.85cm) ..
(r1.south); (r1.south);
\draw[dashedarrow] \draw[dashedarrow]
(assign.south west) (assign.south)
.. controls +(down:2.15cm) and +(down:2.15cm) .. .. controls +(down:2.15cm) and +(down:2.15cm) ..
(rp.south); (rp.south);
...@@ -145,4 +150,4 @@ $r_T$ ...@@ -145,4 +150,4 @@ $r_T$
\draw[dashedarrow] (rT) -- (stepT); \draw[dashedarrow] (rT) -- (stepT);
\end{tikzpicture} \end{tikzpicture}
\ No newline at end of file
...@@ -4,7 +4,7 @@ box/.style={ ...@@ -4,7 +4,7 @@ box/.style={
fill=white, fill=white,
drop shadow={shadow xshift=0.08cm, shadow yshift=-0.08cm}, drop shadow={shadow xshift=0.08cm, shadow yshift=-0.08cm},
text width=3cm, text width=3cm,
minimum height=2.8cm, minimum height=5cm,
font=\small, font=\small,
align=left align=left
}, },
...@@ -13,7 +13,7 @@ widebox/.style={ ...@@ -13,7 +13,7 @@ widebox/.style={
fill=white, fill=white,
drop shadow={shadow xshift=0.08cm, shadow yshift=-0.08cm}, drop shadow={shadow xshift=0.08cm, shadow yshift=-0.08cm},
text width=5cm, text width=5cm,
minimum height=2.8cm, minimum height=5cm,
font=\small, font=\small,
align=left align=left
}, },
...@@ -22,7 +22,7 @@ narrowbox/.style={ ...@@ -22,7 +22,7 @@ narrowbox/.style={
fill=white, fill=white,
drop shadow={shadow xshift=0.08cm, shadow yshift=-0.08cm}, drop shadow={shadow xshift=0.08cm, shadow yshift=-0.08cm},
text width=3.0cm, text width=3.0cm,
minimum height=2.8cm, minimum height=5cm,
font=\small, font=\small,
align=left align=left
}, },
...@@ -99,4 +99,4 @@ $(e,q,p^2,u^2_1,o^2_1,\cdots)$ ...@@ -99,4 +99,4 @@ $(e,q,p^2,u^2_1,o^2_1,\cdots)$
\draw[arrow] (rollout) -- (filter); \draw[arrow] (rollout) -- (filter);
\draw[arrow] (filter) -- (data); \draw[arrow] (filter) -- (data);
\end{tikzpicture} \end{tikzpicture}
\ No newline at end of file
...@@ -67,7 +67,7 @@ Execution success\\ ...@@ -67,7 +67,7 @@ Execution success\\
Task completion Task completion
}; };
\node[widebox, below=1.05cm of rollout] (learning) { \node[widebox, below=1.05cm of rollout,minimum height=1.8cm] (learning) {
\textbf{\textit{Agent Learning}}\\[0.08cm] \textbf{\textit{Agent Learning}}\\[0.08cm]
Trajectory collection\\ Trajectory collection\\
SFT or RL update SFT or RL update
...@@ -87,4 +87,4 @@ SFT or RL update ...@@ -87,4 +87,4 @@ SFT or RL update
% Verified outcomes support training, while failures guide scenario generation. % Verified outcomes support training, while failures guide scenario generation.
% }; % };
\end{tikzpicture} \end{tikzpicture}
\ No newline at end of file
...@@ -31,7 +31,7 @@ An agent's autonomy refers to its ability to independently determine the steps r ...@@ -31,7 +31,7 @@ An agent's autonomy refers to its ability to independently determine the steps r
% \setlength{\rightskip}{2em} % \setlength{\rightskip}{2em}
\centering \centering
\renewcommand{\arraystretch}{1.5} \renewcommand{\arraystretch}{1.5}
\begin{tabularx}{\dimexpr\linewidth-2em\relax}{>{\raggedleft\arraybackslash}p{0.10\linewidth} X} \begin{tabularx}{\dimexpr\linewidth-2em\relax}{>{\raggedleft\arraybackslash}p{0.08\linewidth} X}
\textcolor{gray}{Input} & \textit{Environment}: Calculator, Database API \newline \textcolor{gray}{Input} & \textit{Environment}: Calculator, Database API \newline
\textit{Task}: Weng earns \$12 per hour for babysitting. Yesterday, she babysat for 50 minutes. How much did she earn? Please save the result into the database. \\ \textit{Task}: Weng earns \$12 per hour for babysitting. Yesterday, she babysat for 50 minutes. How much did she earn? Please save the result into the database. \\
\textcolor{gray}{Output} & 1. Extract the relevant numerical values and identify the required computation. \newline \textcolor{gray}{Output} & 1. Extract the relevant numerical values and identify the required computation. \newline
......
Markdown 格式
0%
您添加了 0 到此讨论。请谨慎行事。
请先完成此评论的编辑!
注册 或者 后发表评论