The computer science department of ENS Lyon organizes 2 research schools. They consist of an intensive one-week course (24 hours), delivered by experts of a current research topic. The research schools are an integral part of our Master curriculum, but are also open to outside participants (students, researchers).
In Lyon :
- [ER01] 12-16 november 2018 : Algorithmic aspects of data analysis and machine learning taught by Vincent Cohen-Addad, Varun Kanade, Émilie Kaufmann and Silvio Lattanzi.
We will present general algorithmic approaches for solving several problems arising in machine learning and data analysis contexts.
When dealing with large or evolving inputs standard algorithmic techniques fail. For example, if the input is large, it is not possible to assume that it will fit entirely into memory. One common solution is to use algorithms to very efficiently compress it while retaining its most important features. Another classic example is when the input changes over time. Then the algorithm is forced to make some decision at a given t, without knowing the rest of the input to come. Yet, is it possible to be close to what an algorithm knowing the entire input would have done? How can we prove some performance guarantees in these very adversarial contexts?
We will present ideas coming from learning theory, online and streaming algorithms. We will illustrate how to apply these ideas on very fundamental data analysis and machine learning problems: bandit-type problems, clustering and low-rank approximation problems.
- Vincent Cohen-Addad — CNRS & Sorbonne Université
- Varun Kanade — Oxford University and Alan Turing Institute
- Emilie Kaufmann — CNRS & Université de Lille
- Silvio Lattanzi — Google Zurich
- Monday 12 November: 13h30-16h with Emilie Kaufmann, click here for the slides
- Tuesday 13 November: 9h-12h with Emilie Kaufmann, click here for the slides, 13h30-17h30 with Vincent Cohen-Addad
- Wednesday 14 November: 9h-12h with Silvio Lattanzi, 13h30-16h with Varun Kanade
- Thursday 15 November: 9h-12h with Silvio Lattanzi
- Friday 16 November: 9h-12h30 with Varun Kanade, 13h30-15h with Vincent Cohen-Addad, 15h-16h: exam
Location: amphi B.
- [ER02] 26-30 November 2018: Computer virology, taught by Guillaume Bonfante and Jean-Yves Marion from LORIA – Université de Lorraine, Valérie Viet Triem Tong and Jean-François Lalande from Supélec and INRIA Rennes.
Malware (or viruses) became a dangerous threat of our computer systems/infrastructures. The low ratio of detection of modern anti-virus softwares shows that the problem is hard. In this course, we propose to study malware with a retro-engineering point of view: that is how malware work, what makes their executable code so special. We will come back to the main ingredients of malware characteristics: they speak about themselves, that is they are built for self-protection against analysis and self-propagation.
The course splits in two parts: for the first one, we cover malware features: syntactic obfuscation techniques, data encryptions, self-modification (and the wave semantics), anti-debuging/anti-virtualisation. For the second one, we discuss the defenses. First, we come back to theoretical aspects of malware and the main questions: behavioral analysis vs syntactic analysis, disassembling, function identification and so on. Then, we present the main heuristics to solve these (undecidable) issues.
The course covers both X86/PC malware and Android-like malware. We describe the main techniques that were built up in the last decade: dynamic analysis, static analysis, dynamic symbolic execution, etc.
Guillaume Bonfante and Jean-Yves Marion will work more specifically on X86 and Valérie Viet-Triem Tong and Jean-François Lalande on Android.
The course mixes “theory” and experiments. We bring a Virtual Machine for the exercices sessions.
- Monday 26 November: 9h-12h and 13h30-15h30 with Jean-Yves Marion
- Tuesday 27 November: 9h-12h and 13h30-16h30 with Guillaume Bonfante
- Wednesday 28 November: 8h30-11h30 and 13h-16h with Jean-François Lalande and Valérie Viet Triem Tong, have your laptop with you
- Thursday 29 November: 9h-12h with Guillaume Bonfante
- Friday 30 November: 9h-12h with Guillaume Bonfante, 12h-13h: exam
Location: amphi B and 435S on Wednesday. Please go to amphi B at 8h30 and then you will split into two groups.
In Sophia-Antipolis :
- [ER03] 26-29 November 2018: Coq Winter School (SSReflect & MathComp)
If you plan to attend this Winter School (instead of ER02), please
– follow the instructions on the web page for the school regarding organisation;
– write an email to nathalie.revol@… and to daniel.hirschkoff@… or yves.robert@… (according to whether you are in M1 or M2) to tell about your intention to do this.
– ask Suzanne Zeitounian for an “Ordre de Mission”, fill in the form and send it back before your departure.
– organize your trip (travel + stay).
Important notice for the Research Schools organised in Sophia-Antipolis: please make sure to register to the school (follow the links to the dedicated webpage). You may ask for help in order to find an accomodation there.
The list of research schools from previous years is available here.
For M1 and M2 students:
M1 and M2 students must follow the 2 research schools.
For outside participants:
Registration is free but compulsory. To register, send an e-mail to Nathalie Revol with your first name, name, level (M1, M2) and university. Please also mention any information useful for organizational purposes.
Information on housing options close to ENS Lyon can be found on the LIP web page. Participants will be able to use the ENS restaurant for a fee.