Created
May 15, 2021 16:21
-
-
Save JosephKJ/6609d67e96b99717ed2f0fae0a23b2b1 to your computer and use it in GitHub Desktop.
Algorithm in Latex
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
# Add in the preamble | |
\usepackage{algorithm} | |
\usepackage[noend]{algpseudocode} | |
\renewcommand{\algorithmicrequire}{\textbf{Input:}} | |
\renewcommand{\algorithmicensure}{\textbf{Output:}} | |
# Content | |
\begin{algorithm} | |
\caption{\footnotesize \method \textsc{Inference}} | |
\label{algo:Inference} | |
\begin{algorithmic}[1] | |
\footnotesize | |
\Require{Decoder: $p_{\boldsymbol \theta}(\boldsymbol \psi|\boldsymbol z, \boldsymbol t)$; Last seen task: $\boldsymbol \tau_k$; Task priors: $\mathcal{\boldsymbol P} = \{\boldsymbol P_i\}_{i=1}^{k}$, $\boldsymbol P_i = (\boldsymbol \mu_i, \boldsymbol \Sigma_i)$; Exemplars: $\mathcal{E} = \{Ex_i\}_{i=1}^{m}$, $Ex_i = \{(\boldsymbol x_i, \boldsymbol y_i)\}$; Number of base models to ensemble from: $E$ } | |
% \Ensure{Consolidated Encoder and Decoder parameters: $\phi$ and $\theta$} | |
\If {\textit{Task-agnostic inference}} \Comment{\textit{Task-agnostic inference}} | |
\State $\boldsymbol z \sim \mathcal{N} (\boldsymbol z | \boldsymbol \mu, \boldsymbol \Sigma) $ where $\boldsymbol \mu \gets \frac{1}{k}\sum_{i=1}^{k} \boldsymbol \mu_i$ and $\boldsymbol \Sigma \gets \frac{1}{k}\sum_{i=1}^{k} \boldsymbol \Sigma_i$ | |
\State $\boldsymbol \Psi \gets $ Sample $E$ models from $p_{\boldsymbol \theta}(\boldsymbol \psi | \boldsymbol z)$ | |
\State $\boldsymbol \Psi \gets $ Fine-tune $\boldsymbol \Psi$ on $\mathcal{E}$ | |
\State Ensemble results from $\boldsymbol \Psi$ to solve all tasks ($\boldsymbol \tau_1,\cdots,\boldsymbol \tau_k$) | |
\EndIf | |
\If {\textit{Task-aware inference}} \Comment{\textit{Task-aware inference}} | |
\For{j = 1 to k} | |
\State $\boldsymbol z_j \sim \mathcal{N}(\boldsymbol z|\boldsymbol \mu_j, \boldsymbol \Sigma_j)$ where $\boldsymbol \mu_j, \boldsymbol \Sigma_j \gets \boldsymbol P_j$ | |
\State $\boldsymbol \Psi_j \gets $ Sample $E$ models from $p_{\boldsymbol \theta}(\boldsymbol \psi | \boldsymbol z_j, \boldsymbol t_j)$ | |
\State $\boldsymbol \Psi_j \gets $ Fine-tune $\boldsymbol \Psi_j$ on $Ex_j$ | |
\State Ensemble results from $\boldsymbol \Psi_j$ to solve task $\boldsymbol \tau_j$. | |
\EndFor | |
\EndIf | |
\end{algorithmic} | |
\end{algorithm} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment