Home / Course catalog / Machine Learning Basics for NLP (MC01)


Get this course

expert.ai NL Platform eLearning → Masterclasses

Machine Learning Basics for NLP (MC01)


Description
In this course, you will:
- consolidate your understanding of what machine learning is and of how it works
- review the different types of machine learning: supervised and unsupervised machine learning, as well as semi-supervised machine learning and reinforcement learning
- delve into the lifecycle of a machine learning project, covering the problem framing phase, the data engineering phase, the model selection and tuning phase, the evaluation phase, and the model deployment and monitoring phase
- learn about the main drawbacks of machine learning.
Content
  • Welcome! Course Overview and Objectives
  • I. What is Machine Learning?
  • 1 Machine Learning in a Nutshell
  • 2 Data, Algorithms, and Models
  • 3 The Process and the Assumptions of Machine Learning
  • QUIZ - What is Machine Learning?
  • II. TYPES OF LEARNING
  • 1 Introduction to Types of Learning
  • 2 Supervised Learning
  • 3 Unsupervised Learning
  • 4 Semi-Supervised Learning
  • 5 Reinforcement Learning
  • QUIZ - Types of Learning
  • III. The Machine Learning Lifecycle
  • Introduction to the Machine Learning Lifecycle
  • 1 Problem Framing
  • 1.1 Defining the Performance Targets
  • 1.2 Assessing the Project's Feasibility
  • 1.3 Formulating a Machine Learning Question
  • QUIZ - Problem Framing
  • 2 Data Engineering
  • 2.0 Introduction to Data Engineering
  • 2.1 Raw Data
  • 2.1.1 Data Collection
  • 2.1.2 Data Cleaning
  • 2.1.3 Training Data, Test Data, and Validation
  • 2.1.4 Data Sampling
  • 2.1.5 Detecting Biases in the Data
  • 2.1.6 Data Augmentation
  • QUIZ - Raw Data
  • 2.2 Feature Engineering
  • 2.2.1 Introduction to Feature Engineering
  • 2.2.2 Feature Handcrafting
  • 2.2.3 Feature Extraction
  • 2.2.4 Feature Extraction Methods
  • QUIZ - Feature Engineering
  • 2.3 Vectorization Techniques
  • 2.3.1 Introduction to Text Vectorization Techniques
  • 2.3.2 Bag of Words Vectorization
  • 2.3.3 Tf-idf Vectorization
  • 2.3.4 The Limitations of Bag of Words and Tf-idf
  • 2.3.5 Word Embeddings
  • 2.3.6 Pros and Cons of Word Embeddings
  • QUIZ - Vectorization Techniques
  • 3 Model Selection & Tuning
  • 3.0 Introduction to Model Selection & Tuning
  • 3.1 Model Selection & Tuning: Key Notions
  • 3.1.1 Overfitting and Underfitting
  • 3.1.2 Bias and Variance
  • 3.1.3 Parametric and Non-parametric Algorithms
  • 3.1.4 Parameters and Hyperparameters
  • 3.1.5 How Long to Train a Model?
  • QUIZ - Model Selection & Tuning: Key Notions
  • 3.2 Algorithm Families
  • 3.2.0 Introduction to the Algorithm Families
  • 3.2.1 Regression Algorithms
  • 3.2.1.0 Regression Algorithms
  • 3.2.2 Instance-Based Algorithms
  • 3.2.2.0 Instance-Based Algorithms
  • 3.2.2.1 k-Nearest Neighbors
  • 3.2.2.2 Support Vector Machines
  • 3.2.2.3 Nonlinear Support Vector Machines
  • 3.2.2.4 Pros and Cons of Support Vector Machines
  • 3.2.3 Decision Tree Algorithms
  • 3.2.3.0 Decision Tree Algorithms
  • 3.2.4 Ensemble Algorithms
  • 3.2.4.0 Ensemble Algorithms
  • 3.2.4.1 Bagging and Boosting
  • 3.2.5 Probabilistic Algorithms
  • 3.2.5.0 Probabilistic Algorithms
  • 3.2.5.1 Bayesian Algorithms
  • 3.2.5.2 Naive Bayes
  • 3.2.5.3 Conditional Random Fields
  • 3.2.6 Clustering Algorithms
  • 3.2.6.0 Clustering Algorithms
  • 3.2.7 Association Rule Learning Algorithms
  • 3.2.7.0 Association Rule Learning Algorithms
  • 3.2.8 Artificial Neural Network Algorithms
  • 3.2.8.0 Artificial Neural Network Algorithms
  • 3.2.8.1 How ANNs Are Modeled
  • 3.2.8.2 Comparing different ANNs
  • 3.2.9 Deep Learning Algorithms
  • 3.2.9.0 Deep Learning Algorithms
  • 3.2.9.1 Convolutional Neural Networks
  • 3.2.9.2 Recurrent Neural Networks
  • 3.2.9.3 Transformers
  • QUIZ - Algorithm Families
  • 3.3 Choosing the Right Model
  • 3.3.0 Introduction to Choosing the Right Model
  • 3.3.1 Avoid the State-of-the-Art Trap
  • 3.3.2 Start with the Simplest Model
  • 3.3.3 Avoid Human Biases
  • 3.3.4 Evaluate Performance Over Time
  • 3.3.5 Evaluate Trade-offs
  • 3.3.6 Understand the Model's Assumptions
  • 3.3.7 A Workflow for Selecting a Model
  • 4 Evaluation
  • 4.0 Introduction to Evaluation
  • 4.1 Local Metrics
  • 4.1.1 Model Evaluation and Ground Truth
  • 4.1.2 Building a Confusion Matrix
  • 4.1.3 Accuracy
  • 4.1.4 Precision, Recall, F-Score
  • 4.1.5 Choosing the Right Metrics
  • 4.2 Global Metrics
  • 4.2.1 From Local to Global Evaluation
  • 4.2.2 Working with Imbalanced Datasets
  • 4.2.3 Micro, Macro, and Weighted Metrics
  • 4.2.4 Sample Metrics
  • QUIZ - Evaluation
  • 5 Model Deployment & Performance Monitoring
  • 5.1 Model Deployment
  • 5.2 Performance Monitoring
  • IV. Machine Learning’s drawbacks
  • 1 Introduction to Machine Learning's Drawbacks
  • 2 Data Hunger
  • 3 Lack of Reasoning
  • 4 Poor Reusability
  • 5 Brittleness and Data Shift
  • 6 Opaqueness and Lack of Explainability
  • 7 Computational Costs
  • 8 Reproducibility
  • QUIZ - Machine Learning's Drawbacks
  • Appendix
  • Thank You & Main References
  • Course Feedback Survey
  • TEST - Final Test
Completion rules
  • All units must be completed
  • Leads to a certificate with a duration: 1 year
Back