← All courses
Machine Learning: the fundamentals
Duration : 7 hoursLanguage : EnglishLevel : BeginnerSoftware : A browser; Python is optional, only if you want to reproduce the calculations yourself
Prerequisites:High-school-level algebra (solve a first-degree equation, read a line y = a x + b)Ability to read a simple chart (scatter plot, curve)No programming experience required
Free
The first 2 lessons of this course are open to everyone.
Description
A path into machine learning for people who have never touched it. You start from a scatter plot and a line that fits it, learn to measure a model's error and correct it, then move on to classification (confusion matrix, precision, recall, ROC), neural networks, embeddings and large language models. Every concept rests on a small dataset you can compute by hand, not on a formula you have to take on faith. The course ends with production deployment and fairness, two topics introductory courses too often skip.
Curriculum
The first 2 lessons of this course are open to everyone.
Objectives
- Explain what a linear regression model actually computes and how it learns
- Choose and interpret a loss function (L1, L2, MAE, MSE)
- Read a confusion matrix and compute precision, recall and F1 score by hand
- Understand what a ROC curve is for and what AUC measures
- Tell numerical data apart from categorical data, and spot overfitting
- Explain what neural networks and embeddings are used for
- Describe how a language model predicts the next word, from n-grams to LLMs
- List the questions to ask before shipping a model to production, including on bias
Skills you'll gain
Linear and logistic regressionReading a confusion matrixComputing precision, recall, F1 score, ROC and AUCNeural network and embedding conceptsLanguage model basics (tokens, n-grams, context)Production deployment and algorithmic fairness reference points
Content type
Leçons écrites avec exemples chiffrésTableaux de calcul pas à pasExercices corrigésQuiz interactifs
Your instructor
Équipe OmniLearn