fr
en
Mathematical and statistical foundations for AI
AI model choices become clearer when you understand their mathematical mechanisms. Connect statistics, probability and calculation to modelling situations. Develop a framework for interpreting results and discussing technical choices more precisely.
- Duration
- 1 day 7 hours
- Code
- IA039FR Code
Presentation
Artificial intelligence (AI) relies on strong mathematical foundations that are often perceived as a barrier to entry. This one-day course demystifies these essential concepts and gives you the understanding needed to see what happens under the hood of algorithms.
The programme revisits linear algebra fundamentals, including vectors and matrices, and calculus concepts such as derivatives and gradients, connecting them directly to their role in neural networks. You will also explore descriptive and inferential statistics to analyse data rigorously and interpret the reliability of your predictive models.
The approach is firmly pragmatic: abstract theory is immediately applied. Through workshops using Python and scientific libraries such as NumPy, Pandas and Matplotlib, you will work directly with mathematical concepts to consolidate your understanding and approach future data projects confidently.
Objectives
By the end of this course, you will be able to:
- identify key mathematical concepts, including algebra and probability, used in AI algorithms;
- apply descriptive statistics to analyse and visualise data distributions;
- use inferential statistics to validate hypotheses and improve result reliability;
- interpret a machine learning model's performance metrics and biases, including overfitting;
- use Python tools, including NumPy and Scikit-learn, to explore data and test models.
Program
Module 1: Mastering linear algebra and optimisation basics
- Vectors and matrices: fundamental operations and applications to neural networks.
- Common AI functions, including sigmoid and ReLU, and their derivatives.
- The principle of gradient descent optimisation for model training.
Hands-on exercises
- Manipulate matrices and visualise mathematical functions with NumPy and Matplotlib.
Module 2: Analysing data with statistics and probability
- Measures of central tendency, including mean and median, and dispersion, including standard deviation and variance.
- Data visualisation: interpreting histograms and box plots.
- Common probability distributions, including normal and binomial distributions, and their role in modelling.
Hands-on exercises
- Perform exploratory analysis of a dataset with Pandas and simulate statistical distributions.
Module 3: Validating models through inferential statistics
- Hypothesis tests, including Student's t-test and chi-squared tests, to validate data relevance.
- Interpreting evaluation metrics, including precision, recall and ROC, and managing overfitting.
- Cross-validation to ensure robust predictions.
Hands-on exercises
- Conduct a hypothesis test on a real case, such as Titanic, and evaluate a classification model with Scikit-learn.
Audience
This course is intended for technical professionals seeking to consolidate theoretical foundations, including:
- developers and engineers new to AI who want to understand the mathematical logic behind their libraries;
- junior data analysts and data scientists seeking greater statistical rigour when interpreting results;
- professionals transitioning into data roles who need a targeted mathematics refresher;
- computer science students seeking to connect mathematical theory with machine learning practice.
Prerequisites
The following prerequisites apply:
- Technical knowledge: basic computing and Python programming knowledge is recommended for the workshops.
- Theoretical knowledge: elementary mathematics at secondary-school level.
Teaching and assessment methods
- Initial skills assessment
- Training materials provided to participants
- Continuous assessment throughout the course
- End-of-course feedback questionnaire
- Combination of theory and practical application
- Attendance records
- Post-course follow-up evaluation
- Practical exercises
Course highlights
- Applied learning: abstract mathematical concepts are immediately translated into executable Python code for practical understanding.
- Essential foundations: acquire the background needed to read scientific papers and technical documentation without mathematical barriers.
- AI focus: the programme covers only mathematics relevant to AI, removing unnecessary material to maximise a single day's learning.
- Practical toolkit: leave with reusable code examples for statistical analysis and visualisation.
Dates and sessions
Choose the date and delivery format that suit you.
No upcoming sessions are currently available.
Any brand names and logos mentioned in this course description, such as Python, NumPy, Pandas and Scikit-learn, belong to their respective owners. Their mention for educational purposes does not constitute an endorsement or partnership.