3A / M1 Herbrand (DIX) (2018)

In charge: Sylvie PUTOT, Olivier BOURNEZ

Computers are present in all professional activities that Polytechnicians will have: from management of companies to the development of research or industrial products. In their activities, they will use software tools and will develop new ones. The courses offered to third and fourth year students at the Department of Informatics - Master level - aim to enable students to acquire the knowledge that is necessary to understand the technological changes in the broad field of Science and Technologies of Information and Communication that constitute the landscape in which engineers, civil servants and researchers evolve.

Objectives:

The Computer Science M1-programme (PA Info, or Paris-Saclay M1 Jacques Herbrand) has a triple mission: to provide the students with the core concepts of Computer Science, training them at the level of the best universities in the world; to start a specialization which will continue in fourth year either as a engineering-based training or as a research-based training; and to open up the students perspectives by exploring the fruitful interfaces with related disciplines (from other departments) This will provide the students with a high-level expertise to overcome challenges in industry, administration and research.

Prerequisites:

The generic pre-requisites for a registration in PA Info are to have followed at least to Computer Science courses (excluding modal) in 2nd year, and to have validated a Computer Science project in 2nd year.

Organization:

The PA Info is organized in 10 thematic tracks, 5 of them are centered on Computer Science, 5 offer bi-disciplinary tracks with mathematics. A project 3A has to be taken additionnally to 3 courses per periode. The rules for what form the project 3A can take, depend on tracks. For the tracks Algorithmics and Efficiency and Image Vision, Machine Learning, it takes the form of one additional course or EA per period. For the other tracks, unless explicitely stated by track coordinators, it can either take the form of one additional course or EA per period, or a long research projet. Each track has its own project module, and students who choose a research project must register to the project of their track, even though they may choose a subject advertized on the project module from another track.

Events:

Program INFO

Program MAP INFO

Program MAT INFO

Program INFO

Design of Computer Systems

In charge: Eric GOUBAULT, Sylvie PUTOT

This program is partially supported by the academic chair Engineering of Complex Systems, in partnership between Ecole Polytechnique, Télécom ParisTech, ENSTA ParisTech, and Thalès-DassaultAviation-DCNS-DGA. The chair can offer a certain number of grants for the best internship projects to go to other international institutions (some students went to CMU, GeorgiaTech, Berkely, Austin the last few years).

Objectives:

The « Conception des Systèmes Informatiques » (Design of Computer Systems) track gives the theoretical basis for developing embedded and cyberphysical systems. These systems are blooming, and are already present in all aspects of our daily life : smart-apps on smartphones, primary flight computers (on airplanes) ; autonomous vehicles (Tesla, Google etc.) and drones, possibly organized in swarms ; the smart-grid, among other exciting recent applications. There is no system nowadays, that can be designed, specified, developed, simulated nor verified without computer science through its most fundamental concepts : domain-specific langages, specification langages, formal methods, networks, distributed systems, learning etc. Jobs in these domains range from development to research (there are difficult theoretical problems in these fields), from engineering to project management in big firms or in startup or hi-tech companies.

Leading to:

This tracks has, as natural follow up in 4A : - most of the computer science « mention » tracks of Paris-Saclay, in particular COMASIC (Ecole polytechnique is the coordinator), but also ACN, FIIL etc. or some tracks from the « complex systems engineering » track - most masters in computer science, embedded systems, systems, control, robotics etc., or cyberphysical systems, for example, Master of Computer Science at TUM, CMU, Berkeley, EPFL, ETHZ, KTH Embedded Systems, Systems Controls and Robotics, ETHZ Robotics, Systems and Control, University of California at Irvine or Delft cyberphysical systems etc.

Period 1:

Period 2:

Project: INF510

Internship: INF591


The Internet of Everything For a Digitized Society (IOE4DS)

In charge: Thomas CLAUSEN

This program is partially supported by the academic chair Internet-of-Everything, in partnership between Ecole Polytechnique and Cisco Systems.

Objectives:

In the 21th century, the need for ubiquitous connectivity of everything is unquestionable, for example for: -directly extracting Big Data from the environment, for analysis and decision-making purposes, - remotely piloting and monitoring technical installations for safety, environmental, and social/economical objectives - observing consumption and optimising operation of critical infrastructures, e.g. the next-generation power distribution grid integrating renewable energy, decentralised production, etc., - intelligent living, connected healthcare, smart cities and transport, … Going from internet-scale connectivity of just 10 billion devices today, to ubiquitous connectivity of 50+ billion devices (most of which will be embedded into larger systems) in 10 years, is at the same time an incredible vector of innovation, and a huge technical challenge. This track covers all aspects of that challenge: from the fundamentals of communication, through internet-scale systems engineering, and towards the technologies that in the near (and, not so near) future will allow this ubiquitous connectivity to fulfil its potential and handle the security and privacy issues that come with it - and, of course, to the theoretical foundations enabling this digitisation revolution.

Leading to:

Recommended (well-known) 4A's include Advanced Communication Networks (X-Telecom-ParisTech Paris-Saclay Masters), Stanford CS or EE, KEIO University program in Cyberinformatics

Period 1:

Period 2:

Project: INF512

Internship: INF591


Cyber-Security

In charge: Francois MORAIN

Objectives:

The cybersecurity program is concerned with the security of Internet, at large, and with the ever-increasing set of objects from our environment, that it ties together. Past the first era of discovery and the false illusion of new brotherhood-without-the-old-problems-related-to-bad-guys, people are starting to realize that on the Internet, there're as many bad guys as elsewhere -- and also, that there is nothing like a free lunch in securing our communications nowadays. First, there is no global notion of security. Many problems are security problems: use your laptop and protect your own data; enable and protect the access to Internet; be sure that the communications cannot be eavesdropped, nor altered when you access your bank account; be sure to talk to the right person; and be sure that the communications network infrastructure (such as the Internet, the cell phone system) has not been corrupted and subverted, etc. Since there is no unique way of accessing this virtual world, there is no unique way to tackle security issues. But what is clear is that one needs to understand how the systems work, how resistant they are, how one would design new and more secure applications. The same is true when you want to connect machines to other machines via networks that interfer with the physical world (old wires, WiFi, etc.) and on which information (packets) live their own life in an open though inpredictible way.

Leading to:

M2 MPRI (aka AFP); graduate degree in cyber-security

Period 1:

Period 2:

Project: INF520

Internship: INF591


Algorithms and Foundations of Programming Languages

In charge: Stephane LENGRAND, Gilles SCHAEFFER

Objectives:

This track explore alorithms and programs as scientific concepts. Comparing to the three closest tracks: It has less emphasis on engineering than track 'Design of Computer Systems', and does not offer courses such as Computer architecture; compared to track 'Algorithms and efficiency', it has less emphasis on algorithmics and image processing/computer vision, traded in favour of information theory, cryptology, modelling systems and programs; compared to track 'MAT-Info', it is only made of modules from the CS Department, replacing the algebra and number theory courses by a more extensive coverage of computer science domains.

Leading to:

This track is the first part of the M1-M2 package with the same name 'Algorithms and Foundations of Programming Languages', and that takes place in the Paris area. Students from this M1 track also enrolled or were accepted in Masters abroad, such as MIT, CMU, NYU, Oxford, Imperial, etc.

Footnote:

There is a short but compulsory training course INF549 on OCaml programming, at the very beginning of the year (1h30 lecture + 3h practical). It is especially useful for courses INF551-Computational Logic and INF564-Compilation.

Period 1:

Period 2:

Project: INF513

Internship: INF591


Algorithmics and Efficiency

In charge: Luca Castelli Aleardi, Gilles SCHAEFFER

This program focuses on the design of efficient and robust algorithms, making use of advanced algorithmic techniques, and involving mainly two domains: Image Procession/Computer Vision and Optimization/Operational Research.

Objectives:

The main goal of this program is to provide to students the notions and tools which are required to design efficient and robust algorithms for dealing with real-world problems, that involves structured and complex (huge) data. This is a multidisciplinary program, at the interface of several domains including algorithms, image processing, computer vision, optimization, computer graphics, computational geometry and randomized methods. On the one hand, the goal is to generate curiosity and interest for the theoretical aspects of these domains (complexity of algorithms, parallel/distributed paradigms, ...); on the other hand, this track aims to introduce students to a wide variety of problems and algorithmic solutions, ranging from experimental results to theoretical tools, showing their practical interest and efficiency through real-world applications.

Leading to:

The track will naturally lead to the following Master 2 programs (which are organized in partnership with other institutions): Paris-Saclay Master Program Algorithmique et Fondements de la Programmation (AFP) Master Mathematiques / Vision / Apprentissage (MVA) Master Parisien de Recherche Operationnelle (MPRO) Other Masters programs abroad are also possible: EPFL, ETH, Stanford, Carnegie Mellon, Berkeley, McGill, ... Regarding the research internships in 3A, this track offers multiple possibilities both in academia and in industry. A non-exhaustive list of internships done by students in the previous years includes: Ubisoft, Dassault Systèmes, Facebook, Adobe, Apple, Google, Parrot, Disney, Lucas Films, EADS, Thalès, Siemens, CEA, INRIA, ...

Period 1:

Period 2:

Project: INF514

Internship: INF591


Program MAP INFO

Image, Vision and Machine Learning

In charge: Damien ROHMER, Maks OVSJANIKOV, Renaud KERIVEN

This track, proposed jointly with the Graduate Program in “Artificial Intelligence & Advanced Visual Computing” revolves around the following two main themes: Computer Graphics and Computer Vision, on the one hand, and Artificial Intelligence, on the other, while putting special emphasis on learning-based techniques, in general.

Objectives:

Image processing, or more generally analysis, processing and synthesis of visual data in 2D and 3D are now at the heart of many real-world applications. From the entertainment industry (gaming, special effects, virtual and augmented reality) to applications in analyzing large amounts of visual data (scene understanding, robotics, drones, medical imaging, computer-aided design, social networks, etc.), it is now difficult to imagine communication and reasoning in the digital age without visual data.
The main notions presented in this track include: the basics of analyzing 2D images at the pixel level; Computer Vision and recent techniques based on convolutional neural networks; Computer Rendering on modern graphics hardware; Advanced analysis of 3D shapes and their collections, as well as deformation and animation of 3D models; classical and recent techniques in Artificial Intelligence.
All of these topics aim to provide the basis necessary for addressing broad current scientific challenges, which include seamless integration between synthetic and real data, development of “intelligent” tools for vision and graphics; and real-time interaction with 3D data, among many others.

Finally, the elective modules provide the students a chance for in-depth study of specific aspects of graphics and learning, depending on the interest and intended long-term personal plans. They also give an opportunity to gain better background in more general data analysis, and algorithm design.
The modules of this track build upon techniques in efficient design and programming (C++, GPU) as well as tools from applied and pure mathematics (geometry, numerical optimization, probabilistic analysis) while opening the door to developing creative, personal solutions.

Leading to:

This track can typically lead to the following programs in Master 2 level:
- Ecole Polytechnique and Paris-Sacaly: AIAVC: Graduate Degree Artificial Intelligence & Advanced Visual Computing - MVA: Master Mathématiques/Vision/Apprentissage - AFP: Algorithmique et Fondements de la Programmation (/MPRI) - IMA: Traitement Avancé des Images et Vision, Imagerie du vivant - IAD: Intelligence Artificielle et Décision;
- As well as many Master/Graduate programs abroad at universities such as: EPFL, ETH Zurich, Stanford, Carnegie Mellon, Berkeley, McGill, University of British Columbia, etc.

The research internships at the end of the year (3A) are open to a large number of possibilities in Computer Vision, Computer Graphics and Artificial Intelligence. These general areas are studied in both academic and industrial settings. Examples of companies and laboratories, that have hosted students in the past several years include: Ubisoft, Dassault Systèmes, Facebook, Adobe, Apple, Google, Bentley Systems, Parrot, Disney, Lucas Films, EADS, Thales, Siemens, CEA, Inria, etc.

Footnote:

Students of this track can choose their 4 courses (or 3 course+[INF515]) per period from the provided list with the following additional constraint: At the end of the year, you must have attended, at least, to
- 2 images courses selected among [INF573], [INF574], [INF584], [INF585] (following all of them is possible and even recommended)
- 1 machine learning course selected among [INF554], [MAP569], [INF582], [INF581].

Period 1:

Period 2:

Project: INF515

Internship: INF591


Optimisation

In charge: Claudia D'AMBROSIO

Objectives:

The aim of this track is to introduce students to the decision making science. Decision making is a science studied and used in several disciplines like Engineering, Computer Science, Applied Mathematics, and Economics. The proposed courses will provide the theoretical methods and practical tools to be actors in the decision and optimization process of that institutions and companies have to perform daily. The spectrum covered goes from methods tailored for small to large scale optimization problems to methods to deal with big data.

Leading to:

This track naturally leads to the following M2 programs: - M2 Optimization, Paris Saclay - MPRO : Master Parisien de Recherche Operationnelle - M2RO - Recherche Operationnelle (Toulouse) - track Recherche Operationnelle, Combinatoire et Optimisation proposed in 2nd year of Master Informatique - Montreal, Columbia, etc...

Period 1:

Period 2:

Project: INF516

Internship: INF591


High Performance Computing

In charge: Patrick CARRIBAULT

Objectives:

Le calcul haute performance (HPC) est devenu un point de passage obligé tant dans la recherche academique (validation de nouveaux modèles théoriques) que dans l'industrie (conception de nouveaux produits). De plus les technologies de base du HPC (telle que le parallélisme) se retrouvent utilisées maintenant dans nombre de produits à grande diffusion (tablettes ou téléphones mobiles). Cette fillière vise à former des cadres scientifiques de haut niveau à même de maîtriser deux évolutions technologiques majeures : (i) Utilisation systématique du parallélisme (processeur multi-coeur, processeur graphique, supercalculateur, Cloud Computing), (ii) Utilisation de plus en plus importante du HPC dans les secteurs industriels et recherche.

Leading to:

Master CHPS (Saclay); entreprises françaises ou étrangères

Period 1:

Period 2:

Project: INF518

Internship: INF591


Data Science

In charge: Ioana MANOLESCU, Steve OUDOT

Data Science is a new key area for interdisciplinary science involving mathematics, computer science. It involves a number of challenges in both data management and analyses require new approaches to support the big data era. These challenges span issues regarding generation of the data, their preparation and exploration and of course their analysis. Specific challenges include: dealing with highly distributed data sources, data sampling biases, different formats and structures, data integrity, visualization, machine learning and enabling predictions, real-time analysis and decision-making.

Objectives:

Recently the coupling among Artificial Intelligence methods and Bigdata management creates a very challenging environment where science and industry face huge challenges. This is what makes Data Science a uniquely interesting area with synergies to other fundamental sciences and applications to most scientific and industrial domains.

Leading to:

This parcours equips the students with a solid background necessary for advanced M2 studies in the area of AI and Bigdata Management with a large range of real world applications. Natural choices for M2 studies include the Data Sciences, Advanced Information Content and MVA M2 programs offered by Saclay University. Other choices are abundant in prestigious Universities around the world as Data Science has become a significant area of postgraduate and doctoral training.

Footnote:

MAT557 comes as a complement of INF556 and can therefore be selected only if INF556 itself has been selected. Considering the overlap between INF581 and MAP583, only one of them can be selected at a time.

Period 1:

Period 2:

Project: INF517

Internship: INF592


Program MAT INFO

Common track

In charge: Stephane LENGRAND, Stephane BIJAKOWSKI

Objectives:

The MAT-INFO track aims at providing motivated students with a training in both disciplines that allows them to enroll in a second year of Master either in Math or in CS (or in one of the rare bi-disciplinary Masters). Avoiding an artificial combination of themes, it emphasises the interfaces between the disciplines and covers themes that complement each other: Number theory, Algebra, Cryptology, Logic, Foundations of Mathematics and Computer Science, etc.

Leading to:

The advantage of this track is to keep doors open for the second year: Around Paris, a natural continuation on the CS side is Master 'Algorithms and Foundations of Computer Science'; on the Math side, Masters are available at Université Paris-Sud and Université Pierre et Marie Curie; an excellent bi-disciplinary Master exists at Université Denis Diderot, namely 'Logique Mathématique et Fondements de l'Informatique' (LMFI). Abroad, as in the AFP track, some of our alumni went to (or were accepted by) Stanford, MIT, etc, and a rather unique programme continues the bi-disciplinary training: Oxford's 'Mathematics and the Foundations of Computer Science' (MFoCS). Natural careers for which the MAT-INFO track is particularly well-suited include, but are not limited to, a research-based / academic career or a career in national security institutions (ANSSI, etc).

Footnote:

We ask students to reach a satisfying equilibrium between Math and CS, and therefore pick at least 3 Math courses over the two coursework periods. MAT-INFO students are offered a short training course INF549 on OCaml programming, at the very beginning of the year (1h30 lecture + 3h practical). It is especially useful for INF551-Computational Logic.

Period 1:

Period 2:

Project: INF519

Internship: INF591