ol

handwriting_recognition.html

@@ -71,8 +71,10 @@ <h3>CNN</h3>
 									<p>To illustrate the idea of convolution, I made a graph below to show how convolution filter can extract the feature of vertical edges from an image of T-shirt in the Fashion-MNIST dataset. </p>
 									<img src="images/convolution.png" style="width: 80%; height: auto;">
 									<p>As illustrated as below, deep learning neural network, like CNN, is better than traditional neural networks in two ways: </p>
-									<p>1. CNN can use convolutional filters to extract more meaningful features from images. </p>
-									<p>2. CNN can accommodate much more layers of neurons. </p>
+									<ol>
+										<li>CNN can use convolutional filters to extract more meaningful features from images. </li>
+										<li>CNN can accommodate much more layers of neurons. </li>
+									</ol>
 									<img src="images/deep_nn.png" style="width: 80%; height: auto;">
 									<p>To ensure a fair test, I divided the dataset into a training set (70%) and a testing set (30%). Then, I measured the accuracy of the three abovementioned models using the same testing set. </p>
 								<h2>Results</h2>

index.html

@@ -87,7 +87,7 @@ <h2><a href="#">Binary Search<br />
 									</header>
 									<a href="https://github.com/theo-xiao-sg/algorithms_python" class="image fit"><img src="images/b_v_l.png" alt="" /></a>
 									<p>Here I built python codes to compare the performance of binary search to linear search in terms of their respective calculation speed.</p>
-									<ul class="actions special">1
+									<ul class="actions special">
 										<li><a href="binary_search.html" class="button">Project Description</a></li><li><a href="https://github.com/theo-xiao-sg/algorithms_python" target="_blank" class="button">Source code</a></li>
 									</ul>
 								</article>