Debugs DL/ML blog posts

Author: M.Notter

_posts/2023-10-23-01_scikit_simple.md

@@ -5,25 +5,7 @@ date:   2023-10-23 12:00:00
 description: Getting started with machine learning using Scikit-learn's classification tools
 ---
 
-This post is part of a comprehensive machine learning series that takes you from basic classification to advanced neural networks. Throughout these tutorials, you'll learn:
-
-1. **Getting Started with Machine Learning** (Current Post)
-   Basic classification using Scikit-learn with the MNIST dataset
-   ([View code]({{ site.baseurl }}/assets/scripts/01_scikit_simple.py))
-
-2. **Deep Learning Fundamentals**
-   Introduction to neural networks using TensorFlow
-   ([View code]({{ site.baseurl }}/assets/scripts/02_tensorflow_simple.py))
-
-3. **Advanced Machine Learning**
-   Complex regression pipelines with Scikit-learn
-   ([View code]({{ site.baseurl }}/assets/scripts/03_scikit_advanced.py))
-
-4. **Advanced Deep Learning**
-   Sophisticated neural network architectures in TensorFlow
-   ([View code]({{ site.baseurl }}/assets/scripts/04_tensorflow_advanced.py))
-
-Each tutorial builds upon concepts from previous posts while introducing new techniques and best practices. Whether you're new to machine learning or looking to expand your skills, this series provides hands-on experience with real-world datasets and modern ML tools.
+This post is part of a comprehensive machine learning series that takes you from basic classification to advanced neural networks. Throughout these tutorials, you'll learn machine learning fundamentals using hands-on experience with real-world datasets and modern ML tools.
 
 Have you ever wondered how to get started with machine learning? This series of posts will guide you through practical implementations using two of Python's most popular frameworks: Scikit-learn and TensorFlow. Whether you're a beginner looking to understand the basics or an experienced developer wanting to refresh your knowledge, we'll progress from basic classification tasks to more advanced regression problems.
 

_posts/2023-10-23-02_tensorflow_simple.md

@@ -84,7 +84,7 @@ For image classification, we'll use a Convolutional Neural Network (CNN). CNNs a
 
 There are multiple ways to define a model in TensorFlow. Let's explore two common approaches:
 
-### 1. Sequential API
+### 2.1. Sequential API
 The Sequential API is the simplest way to build neural networks - layers are stacked linearly, one after another:
 
 ```python
@@ -116,7 +116,7 @@ model = keras.Sequential(
 )
 ```
 
-### 2. Layer-by-Layer Sequential API
+### 2.2. Layer-by-Layer Sequential API
 For more explicit control, we can separate each layer and activation:
 
 ```python

_posts/2023-10-23-03_scikit_advanced.md

@@ -178,7 +178,7 @@ So let's setup a pipeline that performs these different pre-processing routines:
 of categorical data to numerical data, data imputer for missing values, data scaling, potential dimensionality
 reduction, etc.
 
-### Handling categorical data
+### 4.1. Handling categorical data
 
 First, let's create a small pipeline that takes categorical data, fills missing values with `'missing'` and
 than applies one-hot encoding on these categorical features.
@@ -201,7 +201,7 @@ categorical_preprocessor = Pipeline(
 )
 ```
 
-### Handling numerical data
+### 4.2. Handling numerical data
 
 To handle numerical data we will use a slightly more advanced processing pipeline (to showcase some scikit-learn
 feature, not because it's the best thing to do). So let's first fill missing values with e.g. the mean of the
@@ -240,7 +240,7 @@ numeric_preprocessor = Pipeline(
 )
 ```
 
-### Combining preprocessing pipelines
+### 4.3. Combining preprocessing pipelines
 
 Now that we have a preprocessing pipeline for the categorical and numerical features, let's combine them into
 one preprocessing pipeline.
@@ -257,7 +257,7 @@ preprocessor = ColumnTransformer(
 )
 ```
 
-### Add regression model
+### 4.4. Add regression model
 
 After the data is preprocessed we want to hand it over to a regression estimator. For this purpose, let's chose
 a ridge regression.

_posts/2023-10-23-04_tensorflow_advanced.md

@@ -649,7 +649,7 @@ results
 
 <img class="img-fluid rounded z-depth-1" src="{{ site.baseurl }}/assets/ex_plots/04_tensorflow_results_table.png" data-zoomable width=800px style="padding-top: 20px; padding-right: 20px; padding-bottom: 20px; padding-left: 20px">
 <div class="caption">
-    Figure 5: Results table for the best performing model architecture, showing both loss and MAE on training and validation sets. The consistent convergence suggests stable learning without overfitting.
+    Figure 3: Results table for the best performing model architecture, showing both loss and MAE on training and validation sets. The consistent convergence suggests stable learning without overfitting.
 </div>
 
 From this table, you could now perform a multitude of follow-up investigations. For example, take a look at the