Retention_Score

Running

App Files Files Community

allenhzy commited on Jun 4, 2024

Commit

eca759a

1 Parent(s): 710289e

equation

Browse files

Files changed (1) hide show

index.html +9 -9

index.html CHANGED Viewed

@@ -437,25 +437,25 @@
       <div class="column container-centered">
         <div id="adaptive-loss-formula" class="container">
           <div id="adaptive-loss-formula-list" class="row align-items-center formula-list">
-            <a href="#label-loss" class="selected">Label Consistency Loss</a>
-            <a href="#representation-loss">Representation Similarity Loss</a>
-            <a href="#total-loss">Total Loss</a>
             <div style="clear: both"></div>
           </div>
           <div id="adaptive-loss-formula-content" class="row align-items-center">
-            <span id="label-loss" class="formula" style="">
               $$
               \displaystyle
               Loss_{label} = \frac{1}{k} \sum_{i=1}^{k} \mathcal{L}\left(\mathbb{C}\left(W^i(x+\delta)  \right), y_t\right)
               $$
             </span>
-            <span id="representation-loss" class="formula" style="display: none;">
               $$
               \displaystyle
               Loss_{repre} = \frac{1}{k} \sum_{i=1}^{k}\mathcal{S}(\mathbb{R}(W^i(x+\delta)), \mathbb{R}(x+\delta))
               $$
             </span>
-            <span id="total-loss" class="formula" style="display: none;">
               $$\displaystyle \mathcal{L}_C(x+\delta, y_t) + Loss_{label} - \alpha \cdot Loss_{repre}$$
             </span>
           </div>
@@ -464,19 +464,19 @@
       <div class="columns is-centered">
         <div class="column">
-          <p id="label-loss" class="eq-des">
             Attackers can design adaptive attacks to try to bypass BEYOND when the attacker knows all the parameters of the model
             and the detection strategy. For an SSL model with a feature extractor $f$, a projector $h$, and a classification head $g$,
             the classification branch can be formulated as $\mathbb{C} = f\circ g$ and the representation branch as $\mathbb{R} = f\circ h$.
             To attack effectively, the adversary must deceive the target model while guaranteeing the label consistency and representation similarity of the SSL model.
           </p>
-          <p id="representation-loss" class="eq-des" style="display: none">
             where $\mathcal{S}$ represents cosine similarity, $k$ represents the number of generated neighbors,
             and the linear augmentation function $W(x)=W(x,p);~p\sim P$ randomly samples $p$ from the parameter distribution $P$ to generate different neighbors.
             Note that we guarantee the generated neighbors are fixed each time by fixing the random seed. The adaptive adversaries perform attacks on the following objective function:
           </p>
-          <p id="total-loss" class="eq-des" style="display: none;">
             where $\mathcal{L}_C$ indicates classifier's loss function, $y_t$ is the targeted class, and $\alpha$ refers to a hyperparameter.
           </p>
         </div>

       <div class="column container-centered">
         <div id="adaptive-loss-formula" class="container">
           <div id="adaptive-loss-formula-list" class="row align-items-center formula-list">
+            <a href=".label-loss" class="selected">Label Consistency Loss</a>
+            <a href=".representation-loss">Representation Similarity Loss</a>
+            <a href=".total-loss">Total Loss</a>
             <div style="clear: both"></div>
           </div>
           <div id="adaptive-loss-formula-content" class="row align-items-center">
+            <span class="formula label-loss" style="">
               $$
               \displaystyle
               Loss_{label} = \frac{1}{k} \sum_{i=1}^{k} \mathcal{L}\left(\mathbb{C}\left(W^i(x+\delta)  \right), y_t\right)
               $$
             </span>
+            <span class="formula representation-loss" style="display: none;">
               $$
               \displaystyle
               Loss_{repre} = \frac{1}{k} \sum_{i=1}^{k}\mathcal{S}(\mathbb{R}(W^i(x+\delta)), \mathbb{R}(x+\delta))
               $$
             </span>
+            <span class="formula total-loss" style="display: none;">
               $$\displaystyle \mathcal{L}_C(x+\delta, y_t) + Loss_{label} - \alpha \cdot Loss_{repre}$$
             </span>
           </div>
       <div class="columns is-centered">
         <div class="column">
+          <p class="eq-des label-loss">
             Attackers can design adaptive attacks to try to bypass BEYOND when the attacker knows all the parameters of the model
             and the detection strategy. For an SSL model with a feature extractor $f$, a projector $h$, and a classification head $g$,
             the classification branch can be formulated as $\mathbb{C} = f\circ g$ and the representation branch as $\mathbb{R} = f\circ h$.
             To attack effectively, the adversary must deceive the target model while guaranteeing the label consistency and representation similarity of the SSL model.
           </p>
+          <p class="eq-des representation-loss" style="display: none">
             where $\mathcal{S}$ represents cosine similarity, $k$ represents the number of generated neighbors,
             and the linear augmentation function $W(x)=W(x,p);~p\sim P$ randomly samples $p$ from the parameter distribution $P$ to generate different neighbors.
             Note that we guarantee the generated neighbors are fixed each time by fixing the random seed. The adaptive adversaries perform attacks on the following objective function:
           </p>
+          <p class="eq-des total-loss" style="display: none;">
             where $\mathcal{L}_C$ indicates classifier's loss function, $y_t$ is the targeted class, and $\alpha$ refers to a hyperparameter.
           </p>
         </div>