Machine Learning for C# Developers Made Easy (The Course)

1.3.1 The main benefits of choosing ML.NET

1.3.2 What makes ML.NET easier to use than alternatives

1.4.1 Building products and services

1.4.2 Automating internal tasks

1.5.1 How ML.NET helps Power BI identify key influencers

1.5.2 Other organizations that rely on ML.NET

1.6.1 Selecting a task

1.6.2 Preparing the data

1.6.3 Training and evaluating the model

1.6.4 Using trained models

1.7.1 Minimal system requirements

1.7.2 Installing ML.NET via the command line

1.7.3 Updating the tool

1.7.4 Installing Model Builder in Visual Studio

2.2.1 Regression

2.2.2 Classification

2.3.1 Clustering

2.3.2 Anomaly detection

2.3.3 Association rule learning

2.4.1 The reinforcement learning process

2.5.1 Training a classification model

2.5.2 Auto-generated ML.NET project

Listing 2.1 ML.NET classification project structure

2.5.3 Retraining the classification model

Listing 2.2 Method for retraining a pipeline

Listing 2.3 Method for building a pipeline

2.5.4 Consuming a trained classification model

Listing 2.4 Model input data object

Listing 2.5 Model output data object

Listing 2.6 Input data object

Listing 2.7 Output data object

Listing 2.8 Preparing model inputs and outputs

Listing 2.9 Building the pipeline

Listing 2.10 Creating input object for a prediction

3.1.1 Deciding between supervised and unsupervised training

3.1.2 Choosing a supervised learning task

3.1.3 Choosing an unsupervised learning task

3.2.1 Binary and multiclass classification

3.2.2 Image classification

3.2.3 Regression

3.2.4 Recommendation

3.2.5 Forecasting

3.2.6 Other supervised learning tasks supported by ML.NET

3.3.1 The usage of clustering in ML.NET

3.3.2 The usage of anomaly detection in ML.NET

3.4.1 Quality metrics for supervised learning

3.4.2 Accuracy metrics for unsupervised learning

3.4.3 Accuracy metrics in ML.NET

Listing 4.1 CLI interface placeholder

4.2.1 Training the classification model

4.2.2 Reviewing the classification pipeline

Listing 4.2 Multiclass classification pipeline

4.2.3 Making our model consumable

Class library

Listing 4.3 Tech support classification

4.2.4 Interacting with the model

Listing 4.4 Application start-up code with classification

4.3.1 Training the regression model

4.3.2 Reviewing the regression pipeline

Listing 4.5 Regression training pipeline

4.3.3 Adding the regression capabilities to the virtual assistant

Listing 4.6 Tech support classification

Listing 4.7 Application start-up code with a regression task

4.4.1 Training the recommendation model

4.4.2 Reviewing the recommendation pipeline

Listing 4.8 Recommendation pipeline

4.4.3 Adding the recommendation capabilities to the virtual assistant

Listing 4.9 Recommendation logic

Listing 4.10 Application startup code with recommendation

4.5.1 Training the forecasting model

4.5.2 Reviewing the forecasting pipeline

Listing 4.11 Forecasting training pipeline

4.5.3 Enabling forecasting in the virtual assistant app

Listing 4.12 Forecasting consumption logic

Listing 4.13 Complete application startup logic

Listing 5.1 Input data item

Listing 5.2 Output data item

Listing 5.3 data transfer object with input data

5.1.1 Adding clustering code

Listing 5.4 Cluster data storage

5.1.2 Adding the model consumption code

Listing 5.5 Cluster prediction class

5.1.3 Adding the model to the virtual assistant

Listing 5.6 Adding the predictor to the virtual assistant

Listing 5.7 Application start-up code with clustering

5.1.4 Testing clustering functionality

Listing 5.8 Anomaly detection input data item

Listing 5.9 Anomaly detection output

5.2.1 Adding anomaly detection code

Listing 5.10 The entry method for anomaly detection

5.2.2 Adding anomaly detector to the virtual assistant

Listing 5.11 Anomaly detection consumption logic

Listing 5.12 Complete application entry point logic

5.2.3 Testing our anomaly detection logic

5.4.1 Building a real estate categorization model

Listing 5.13 Input data object for clustering model

5.4.2 Building a code smells categorization model

6.2.1 Training a sentence similarity model

Listing 6.1 Input data for sentence similarity

Listing 6.2 Output data for sentence similarity

Listing 6.3 Training pipeline for sentence similarity

6.2.2 Consuming the sentence similarity model

Listing 6.4 Prediction engine for sentence similarity

6.2.3 Adding sentence similarity checker to the virtual assistant

Listing 6.5 Sentence Comparer helper class

6.2.4 Pre-processing sentence similarity data

6.2.5 extracting features from text tokens

6.2.6 Determining the similarity between two sentences

Listing 6.6 Chatbot input processor

6.4.1 Hallucinations in AI and where they come from

6.5.1 Training a NER model with ML.NET

Listing 6.7 NER input class

Listing 6.8 NER output class

Listing 6.9 NER label class

Listing 6.10 seeding label data

Listing 6.11 training data processor

Listing 6.12 Application entry point

6.5.2 Using the NER model

6.5.3 How NER works

6.5.4 Practical applications of NER

6.5.5 An example of a NER training model

6.6.1 Enhancing the training data

6.6.2 Guarding against hallucinations

Listing 6.13 Text classification output class

6.6.3 Incorporating NER in our chatbot

7.1.1 A brief overview of TensorFlow

7.1.2 Training pipeline for image classification

Listing 7.1 Image classification pipeline

Listing 7.2 Loading image data from folders

Listing 7.3 Image classification input object

Listing 7.4 Image classification output object

7.1.3 Consuming image classification model

Listing 7.5 Console interface for image

7.2.1 Training pipeline for object detection

Listing 7.6 Object detection training pipeline

Listing 7.7 Object detection training process

Listing 7.8 Object detection input model

Listing 7.9 Object detection output model

7.2.2 Consuming an object detection model

Listing 7.10 Using a trained object detection model

7.3.1 Adding TensorFlow integration

Listing 7.11 TensorFlow Inception settings

Listing 7.12 Input data

Listing 7.13 Prediction data structure

7.4.1 Image classification: recognizing what’s in an image

7.4.2 Object detection: locating and identifying objects

8.1.1 Adding a model to an existing business workflow

Listing 8.1 Refund processor workflow orchestrator function

8.1.2 System bottlenecks

8.1.3 Training the model

8.1.4 Using Model Builder to enhance human-in-the-loop flow

8.2.1 Computer vision

8.2.2 Natural language processing

9.1.1 Initial setup

9.1.2 Loading the training data

9.1.3 Obtaining the embeddings

9.1.4 Building and configuring the training pipeline

9.1.5 Running AutoML experiments

9.1.6 Evaluating results

9.2.1 A practical AutoML pattern you can reuse

Listing 9.1 Reusable AutoML template

9.2.2 Regression with AutoML

Listing 9.2 Regression pipeline in AutoML

9.2.3 Binary classification with AutoML

Listing 9.3 Binary classification pipleine in AutoML

9.2.4 Multiclass classification with AutoML

Listing 9.4 Multiclass classification pipleine in AutoML

9.2.5 Recommendation with AutoML (and the common fallback)

Listing 9.5 Recommendation pipleine in AutoML

Listing 9.6 Recommendation pipleine in AutoML with Matrix Factorization trainer

9.3.1 Tuning the sweepable pipeline: include or exclude trainers

Listing 9.7 Including and excluding trainers

9.3.2 Tuning the search space: change the range AutoML sweeps

Listing 9.8 Configuring the search space

9.3.3 Creating a custom sweepable estimator

Listing 9.9 Custom sweepable estimator

9.3.4 Combining default pipelines with custom sweepable estimators

Listing 9.10 Default pipeline combined with a sweepable estimator

9.3.5 When tuning is worth the effort

10.1.1 Importing an ONNX model into ML.NET

Listing 10.1 Importing ONNX model into the ML.NET pipeline

10.1.2 Exporting an ML.NET model to ONNX

Listing 10.2 Exporting an ML.NET model as ONNX

Listing 10.3 Importing a TensorFlow model into ML.NET

10.2.1 Exporting an ML.NET model to TensorFlow

10.3.1 Exporting an ML.NET model to PyTorch

10.4.1 The caveats that save you time

10.5.1 ONNX

10.5.2 TensorFlow

PyTorch