Update index.html

talks/cawl-clustering/index.html

@@ -237,7 +237,7 @@ <h3>Variational Auto-Encoder (VAE)</h3>
                                 3 stacked convolutional layers
                             </li>
                             <li class='fragment semi-fade-in'>
-                                Trained to reconstruct character images from dense bottleneck $z \sim \mathcal{N}(\mu, \sigma)$
+                                Trained to reconstruct character images from bottleneck $z \sim \mathcal{N}(\mu, \sigma)$
                             </li>
                             <!--<li class='fragment semi-fade-in'>
                                 Train with the same soft-clustering pseudolabel objective used by Sign2Vec