Programme
Three years, six semesters, 180 ECTS credits. The guiding principle: to work seriously on artificial intelligence you need the mathematics and the computer science underneath it, and some understanding of the world the systems are deployed into.
The shape of the three years
Each year carries 60 ECTS credits, divided into two semesters of around fourteen teaching weeks. Courses are scheduled Monday to Friday, on average 25 hours a week of lectures and tutorials, not counting personal work.
The first year is mostly the fundamentals of mathematics and computer science, with artificial intelligence introduced alongside them; the second keeps that mathematical core going but gives most of the new ground to the methods of current AI; the third is for advanced topics, for specialization in the fields AI is applied to, for a data challenge, and for an internship. Every course is listed on the page of its year: year 1, year 2, year 3.
Three blocks of courses
Every course belongs to one of three blocks, which keeps the curriculum balanced and shapes the way assessment and progression work.
- Foundations of AI
- Mathematics and computer science: algebra, analysis, probability and statistics, programming, algorithms, databases, optimization, measure theory. These carry most of the credits, particularly in the first two years.
- AI-focused courses
- Artificial intelligence itself, from the introductory course in the first semester through machine learning, deep learning and reinforcement learning to the specializations of the third year – and practical work that puts them to use: a data challenge, and the research internship or industrial placement that ends the degree.
- Transverse courses
- Economics, law, philosophy and ethics, cognitive science, the environmental cost of computing, scientific English, sport, and the personal project of the second year. Short courses, present in every semester but the last. Students may also take an optional language course – French as a foreign language, or another language – at the language centre of PSL, which carries no credit.
| Year | Semester | Foundations of AI | AI-focused | Transverse |
|---|---|---|---|---|
| Year 1 | 1 | 24 | 1.5 | 4.5 |
| 2 | 23 | 2.5 | 4.5 | |
| Year 2 | 3 | 24 | 3 | 3 |
| 4 | 19 | 7 | 4 | |
| Year 3draft | 5 | 17 | 11 | 2 |
| 6 | – | 30 | – |
The shift over the three years is visible in the figures: the foundations carry almost everything at the start and give way, semester by semester, to the courses on artificial intelligence itself.
How it is taught
Courses combine lectures with tutorials and lab sessions in small groups: for the main mathematics and programming courses, typically three hours of lecture and three hours of tutorial a week. Teaching is in English throughout, by researchers and teaching staff from the PSL schools and laboratories.
Marks come from continuous assessment, written examinations and project work, in proportions that depend on the course. The academic year has two examination periods, with a resit session for students who have not passed a block; see the calendar.
What the degree certifies
The bachelor is a degree of Université PSL (a diplôme de grand établissement) that confers the grade de licence, the French bachelor’s level. It is registered as RNCP42263, which sets out the competencies it certifies: five blocks common to every French bachelor’s degree – digital tools, working with data, communicating in a foreign language, situating oneself in a professional field, working responsibly within an organization – and two specific to this one. Among the latter, a graduate is certified to be able to:
- Analyse a problem and turn it into functional and technical specifications for an AI solution.
- Collect, clean and prepare datasets for machine learning, documenting the choices made in preprocessing.
- Build supervised and unsupervised learning systems fitted to a problem, improving the balance of risks and benefits, environmental and socio-economic, from the design stage onwards.
- Tune deep learning architectures under constraints of performance and of resource consumption.
- Choose evaluation metrics suited to the models, analysing the quality, the biases, and the limits of the data.
- Integrate AI models into simple software applications.
- Identify the main ethical, legal, societal and environmental questions raised by artificial intelligence – bias, discrimination, explainability – in the various contexts of its use, and take them into account in methodological choices.
- Communicate technical results to audiences of different kinds, in speech and in writing, spelling out the ethical and environmental implications.
The full framework, in French, is on the France Compétences record of the diploma.
Where the degree leads
The bachelor is meant primarily as a preparation for graduate study rather than as a terminal professional degree. Within Université PSL, graduates go on to master’s programmes such as Artificial Intelligence, Systems, Data (IASD), in its computer science or mathematics track, or AI & Society; to the research-oriented master’s programme of ENS-PSL, through a competitive admission process; or to a master’s programme in a selective university elsewhere in France or abroad. A doctorate, in computer science or mathematics, follows for those who want to do research, in a PSL laboratory or elsewhere.
The degree also qualifies its holders for work directly, as data analyst, specialist technician in artificial intelligence or in data processing, or developer specialized in artificial intelligence. Two to five years of experience, or further study, lead on from there to data scientist, data engineer, machine learning engineer, AI software engineer, AI consultant, or researcher.
The degree
Graduates receive a diploma awarded by Université PSL, a diplôme de grand établissement, conferring the grade de licence in artificial intelligence: level 6 of the French and European qualification frameworks, worth 180 ECTS credits.
