Stage

What is the common point between the affectation of students to universities (in particular, in
France, the Parcoursup algorithm), the influence of someone in a social network (like Twitter), the
responsibility of a formula in the inconsistency of a belief base, the impact and synergy of some criteria
in a multi-criteria decision making situation? In all these situations we are confronting with a set of
objects/persons on which we have information on the results of some coalitions/groups, and where it is

The aim of this internship is to analyze group interaction and their evolution over time. We will rely on existing data (images and videos) of group interaction that have been annotated at different levels (activity, speaking, laughing, non-verbal behavior). We will first make use of the database MatchNMingle (Raman&Hung,19).

Les avancées technologiques ainsi que l’introduction de intelligence artificielle et des
sciences cognitives dans le monde de l’éducation ont permis des grandes avancées
dans ce domaine.
Dans ce stage, nous souhaitons utiliser l’intelligence artificielle pour analyser les traces
des mouvements oculaires (oculométrie) de personnes en situation d’apprentissage
en cherchant à les associer à d’autres données issus de l’analyse de la vidéo ou des
traces d’interaction. Les analyses des données se feront à partir de vidéos et seront

Markov Decision Processes (MDPs) and reinforcement learning (RL) are two very successful paradigms adopted in artificial intelligence for designing autonomous agents capable of dealing with sequential decision problems under uncertainty. The decision problem is formalized as a tuple <S, A, T, R> with S the set of system states, A the set of possible actions, T the transition function and R the reward function.

Ces travaux s’inscrivent dans le cadre de recherches sur l’autonomie à long terme (mois, années) d’une entité artificielle.
Concevoir des mécanismes non-déportés à même de doter une entité artificielle non-connectée d’un tel niveau d’autonomie nécessite de se poser la question de la représentation et de la mise à jours des connaissances - par nature multimodales - manipulées.

Après avoir identifié les principales approches existantes pour l’apprentissage et la représentation de connaissances dans un contexte multimodal, vous analyserez celles-ci sous l’angle de l’autonomie à long-terme d’un système déconnecté, et de l’approche développementale. Fort de cette analyse, vous proposerez une architecture apprentissage-représentation-exploitation adaptée à l’autonomie à long terme.

Fair division is a fundamental problem raised in many real word problems such as multi-robot task allocation, allocation of schools, courses
or rooms to students, division of goods ininheritance or divorce settlement… A multi agent fair division problem consists in dividing a set of resources among a set of agents while respecting a fairness notion.
Various fairness notions have been proposed and studied in economics and computer science.

Internship proposition to be supervised by Christoph Dürr, LIP6, Sorbonne University

Suppose you need to buy an airline ticket for your next vacations which start in $n$ days. Every day you see a different price. Once you decided to buy the ticket the game is over. How should you decide when to buy? This is the topic of ongoing research on a problem called *prophet inequalities*.

See pdf for details.

Bleuse et al. (EuroPar 2018) [1] introduced a general model for interference-aware scheduling in large scale parallel platforms. They considered two different types of communications: the flows induced by data exchanges during computations and the flows related to Input/Output operations. Rather than taking into account these communications explicitly, they restrict the possible allocations of a job by external topological constraints.

<b>Contexte</b>

to analyze group interaction. We will rely on existing data (images and
videos) of group interaction that have been annotated at different levels (activity, speaking, laughing,
non-verbal behavior). We will first make use of the database MatchNMingle (Raman&Hung,19).
Several steps are foreseen:
1. Perform a literature survey on F-formation detection.
2. Extract regions of interest in static images, compute their descriptors, and classify the
situation in terms of F-formation.
3. Extend the model to videos.