Modern applications, such as social networking systems and e-commerce platforms are centered around using large-scale databases for storing and retrieving data. Accesses to the database are typically enclosed in transactions that allow computations on shared data to be isolated from other concurrent computations and resilient to failures. Modern databases trade isolation for performance. The weaker the isolation level is, the more behaviors a database is allowed to exhibit and it is up to the developer to ensure that their application can tolerate those behaviors.

In this work, we propose stateless model checking algorithms for studying correctness of such applications that rely on dynamic partial order reduction. These algorithms work for a number of widely-used weak isolation levels, including Read Committed, Causal Consistency, Snapshot Isolation and Serializability. We show that they are complete, sound and optimal, and run with polynomial memory consumption in all cases. We report on an implementation of these algorithms in the context of Java Pathfinder applied to a number of challenging applications drawn from the literature of distributed systems and databases.

We will survey a rather recent and very fruitful line of research in algorithmic fairness, coined multi-group fairness. We will focus on risk prediction, where a machine learning algorithm tries to learn a predictor to answer questions of the form “what is the probability that patient x will experience a particular medical condition?” Training a risk predictor to minimize a loss function fixed in advance is the dominant paradigm in machine learning. However, global loss minimization may create predictions that are mis-calibrated on sub-populations, causing harm to individuals of these populations. Multi-group fairness tries to prevent forms of discrimination to a rich (possibly exponential) collection of arbitrarily intersecting groups. In a sense, it provides a computational perspective on the meaning of individual risks and the classical tension between clinical prediction, which uses individual-level traits, and actuarial prediction, which uses group-level traits.

While motivated in fairness, this alternative paradigm for training an indistinguishable predictor is finding a growing number of appealing applications, where the same predictor can later be used to optimize one of a large set of loss functions, under a family of capacity and fairness constraints and instance distributions.

Based on a sequence of works joint with (subsets of) Cynthia Dwork, Shafi Goldwasser, Parikshit Gopalan, Úrsula Hébert-Johnson, Lunjia Hu, Adam Kalai, Christoph Kern, Michael P. Kim, Frauke Kreuter, Guy N. Rothblum, Vatsal Sharan, Udi Wieder, Gal Yona and others.

Secure Multiparty Computation (MPC) aims at enabling different parties to compute functions based on private inputs. Providing the parties with additional correlated and random inputs enables highly efficient MPC protocols. But how do we construct this additional correlated randomness? In this presentation, I will discuss the construction of pseudorandom correlation generators (PCG), focusing on the special case of two-party random oblivious linear evaluations (OLEs). We will explore the foundational concepts introduced by Boyle et al. (2020), examine their limitations, and demonstrate how coding theory significantly contributes to the security analysis and efficiency of the construction. This presentation is based on collaborative work with Maxime Bombar, Dung Bui, Geoffroy Couteau, Alain Couvreur, and Sacha Servan-Schreiber.

We elaborate on a problem raised by Schmidt in 1967: rational approximation of linear subspaces of $\mathbb{R}^n$. In order to study the quality approximation of irrational numbers by rational ones, one can introduce the exponent of irrationality of a number. We can then generalize this notion in the framework of vector subspaces for the approximation of a subspace by so-called rational subspaces. After briefly introducing the tools for constructing this generalization, I will present the different possible studies of this object. Finally I will explain how we can construct spaces with prescribed exponents.

In my talk, I will discuss the category $pOpe_\iota$ of positive opetopes with \iota-contractions, introduced by Zawadowski. I will then define the category $\widehat{pOpe_\iota}_{EZ}$ of Eilenberg-Zilber opetopic sets, which are presheaves on $pOpe_\iota$ analogous to simplicial sets. I will outline a modification of Cisinski theory applicable to $\widehat{pOpe_\iota}_{EZ}$, demonstrating the existence of a model structure. Additionally, I will sketch the proof of its Quillen equivalence to the Kan-Quillen model structure. If time allows, I will touch upon another model structure and its potential comparison with the Joyal model structure on simplicial sets.