diff --git a/lib/scholar/manifold/mds.ex b/lib/scholar/manifold/mds.ex
index 7976c9da..ae62b720 100644
--- a/lib/scholar/manifold/mds.ex
+++ b/lib/scholar/manifold/mds.ex
@@ -1,10 +1,7 @@
 defmodule Scholar.Manifold.MDS do
   @moduledoc """
-  TSNE (t-Distributed Stochastic Neighbor Embedding) is a nonlinear dimensionality reduction technique.
-
-  ## References
-
-  * [t-SNE: t-Distributed Stochastic Neighbor Embedding](http://www.jmlr.org/papers/volume9/vandermaaten08a/vandermaaten08a.pdf)
+  Multidimensional scaling (MDS) seeks a low-dimensional representation of the data
+  in which the distances respect well the distances in the original high-dimensional space.
   """
   import Nx.Defn
   import Scholar.Shared
diff --git a/notebooks/cv_gradient_boosting_tree.livemd b/notebooks/cv_gradient_boosting_tree.livemd
index 9fe6d970..f9a6a970 100644
--- a/notebooks/cv_gradient_boosting_tree.livemd
+++ b/notebooks/cv_gradient_boosting_tree.livemd
@@ -1,4 +1,4 @@
-# Using cross-validation with gradient boosting trees
+# Cross-validation with gradient boosting trees
 
 ```elixir
 Mix.install([
diff --git a/notebooks/hierarchical_clustering.livemd b/notebooks/hierarchical_clustering.livemd
index b7096029..fd86e0d4 100644
--- a/notebooks/hierarchical_clustering.livemd
+++ b/notebooks/hierarchical_clustering.livemd
@@ -1,4 +1,4 @@
-# Hierarchical Clustering
+# Hierarchical clustering
 
 ```elixir
 app_root = Path.join(__DIR__, "..")
diff --git a/notebooks/k_means.livemd b/notebooks/k_means.livemd
index d747d825..a8b22086 100644
--- a/notebooks/k_means.livemd
+++ b/notebooks/k_means.livemd
@@ -1,6 +1,6 @@
 <!-- livebook:{"persist_outputs":true} -->
 
-# K-means clustering
+# k-means clustering
 
 ```elixir
 Mix.install([
diff --git a/notebooks/k_nearest_neighbors.livemd b/notebooks/k_nearest_neighbors.livemd
index 01f27f7d..014e3df5 100644
--- a/notebooks/k_nearest_neighbors.livemd
+++ b/notebooks/k_nearest_neighbors.livemd
@@ -1,6 +1,6 @@
 <!-- livebook:{"persist_outputs":true} -->
 
-# K-Nearest Neighbors
+# k-nearest neighbors
 
 ```elixir
 Mix.install([
diff --git a/notebooks/linear_regression.livemd b/notebooks/linear_regression.livemd
index 81f69e9c..a5c83c26 100644
--- a/notebooks/linear_regression.livemd
+++ b/notebooks/linear_regression.livemd
@@ -1,6 +1,6 @@
 <!-- livebook:{"persist_outputs":true} -->
 
-# Linear Regression in Practice
+# Linear regression in practice
 
 ```elixir
 Mix.install([
diff --git a/notebooks/mds.livemd b/notebooks/mds.livemd
index b47f4bc9..21d3ecd7 100644
--- a/notebooks/mds.livemd
+++ b/notebooks/mds.livemd
@@ -1,4 +1,4 @@
-# MDS
+# Multidimensional scaling (MDS)
 
 ```elixir
 Mix.install([
@@ -25,6 +25,8 @@ alias Explorer.DataFrame, as: DF
 
 ## Swiss Roll
 
+Multidimensional scaling (MDS) seeks a low-dimensional representation of the data in which the distances respect well the distances in the original high-dimensional space.
+
 The first example demonstrating MDS will use so-called Swiss Roll dataset. It's call as such because it resembles a popular sponge cake. Let's have a look on this dataset.
 
 | ![Skewness](https://scikit-learn.org/stable/_images/sphx_glr_plot_swissroll_001.png)                        |
@@ -1098,7 +1100,7 @@ labels
 Let's look how one of the images look like
 
 ```elixir
-image = 
+image =
   Nx.reshape(digits_data[[1234, ..]], {image_height, image_width})
   |> Nx.as_type({:u, 8})