Fundamentals of Artificial Intelligence (AI) and Machine Learning Training Course

Introduction to Applied Machine Learning

• Statistical learning compared to machine learning
• The process of iteration and model evaluation
• The Bias-Variance trade-off
• Supervised versus Unsupervised Learning
• Challenges addressed through Machine Learning
• Train, Validation, and Test splits – The ML workflow for preventing overfitting
• The standard Machine Learning workflow
• Overview of machine learning algorithms
• Selecting the most suitable algorithm for specific problems

Algorithm Assessment

• Assessing numerical forecasts
  • Accuracy metrics: ME, MSE, RMSE, MAPE
  • Stability of parameters and predictions
• Evaluating classification techniques
  • Accuracy and its associated limitations
  • Utilizing the confusion matrix
  • Addressing issues with unbalanced classes
• Visualizing model efficacy
  • Profit curves
  • ROC curves
  • Lift curves
• Model selection strategies
• Model optimization – Grid search techniques

Data Preparation for Modelling

• Importing and storing data
• Initial data exploration and comprehension
• Manipulating data using the pandas library
• Data transformation – Data wrangling
• Conducting exploratory analysis
• Handling missing observations – Detection and remedies
• Outliers – Identification and management strategies
• Standardization, normalization, and binarization
• Recoding qualitative data

Machine learning algorithms for Outlier detection

• Supervised algorithms
  • KNN
  • Ensemble Gradient Boosting
  • SVM
• Unsupervised algorithms
  • Distance-based approaches
  • Density-based methods
  • Probabilistic methods
  • Model-based methods

Understanding Deep Learning

• Fundamental concepts of Deep Learning
• Distinguishing Machine Learning from Deep Learning
• Key applications of Deep Learning

Overview of Neural Networks

• Defining Neural Networks
• Neural Networks compared to Regression Models
• Mathematical foundations and learning mechanisms
• Constructing an Artificial Neural Network
• Neural nodes and their connections
• Managing neurons, layers, and input/output data
• Single Layer Perceptrons explained
• Differences between Supervised and Unsupervised Learning
• Feedforward and Feedback Neural Networks
• Forward Propagation and Back Propagation concepts

Building Simple Deep Learning Models with Keras

• Initializing a Keras Model
• Analyzing and understanding the dataset
• Defining the Deep Learning architecture
• Compiling the Model
• Fitting the Model to data
• Handling Classification Data
• Developing Classification Models
• Deploying and using the Models

Working with TensorFlow for Deep Learning

• Data Preparation
  • Data acquisition
  • Preparing Training Data
  • Preparing Test Data
  • Scaling input features
  • Utilizing Placeholders and Variables
• Defining the Network Architecture
• Implementing the Cost Function
• Applying the Optimizer
• Using Initializers
• Training the Neural Network
• Constructing the Graph
  • Inference phase
  • Loss calculation
  • Training phase
• Training the Model
  • Managing the Graph
  • Utilizing the Session
  • The Train Loop
• Model Evaluation
  • Creating the Evaluation Graph
  • Assessing performance with Eval Output
• Scaling Model Training
• Visualizing and assessing models with TensorBoard

Application of Deep Learning in Anomaly Detection

• Autoencoders
  • Encoder-Decoder Architecture
  • Reconstruction loss metrics
• Variational Autoencoders
  • Variational inference techniques
• Generative Adversarial Networks
  • Generator–Discriminator architecture
  • Anomaly detection approaches using GANs

Ensemble Frameworks

• Aggregating results from diverse methods
• Bootstrap Aggregating (Bagging)
• Averaging outlier scores