Search

Wright’s inbreeding coefficient, FST, is a fundamental measure in population genetics. Assuming a predefined population subdivision, this statistic is classically used to evaluate population structure at a given genomic locus. With large numbers of loci, unsupervised approaches such as principal component analysis (PCA) have, however, become prominent in recent analyses of population structure. In this study, we describe the relationships between Wright’s inbreeding coefficients and PCA for a model of K discrete populations. Our theory provides an equivalent definition of FST based on the decomposition of the genotype matrix into between and within-population matrices. The average value of Wright’s FST over all loci included in the genotype matrix can be obtained from the PCA of the between-population matrix. Assuming that a separation condition is fulfilled and for reasonably large data sets, this value of FST approximates the proportion of genetic variation explained by the first (K − 1) principal components accurately. The new definition of FST is useful for computing inbreeding coefficients from surrogate genotypes, for example, obtained after correction of experimental artifacts or after removing adaptive genetic variation associated with environmental variables. The relationships between inbreeding coefficients and the spectrum of the genotype matrix not only allow interpretations of PCA results in terms of population genetic concepts but extend those concepts to population genetic analyses accounting for temporal, geographical and environmental contexts., Author summary Principal component analysis (PCA) is the most-frequently used approach to describe population genetic structure from large population genomic data sets. In this study, we show that PCA not only estimates ancestries of sampled individuals, but also computes the average value of Wright’s inbreeding coefficient over the loci included in the genotype matrix. Our result shows that inbreeding coefficients and PCA eigenvalues provide equivalent descriptions of population structure. As a consequence, PCA extends the definition of those coefficients beyond the framework of allelic frequencies. We give examples on how FST can be computed from ancient DNA samples for which genotypes are corrected for coverage, and in an ecological genomic example where a proportion of genetic variation is explained by environmental variables.

A major objective of evolutionary biology is to understand the processes by which organisms have adapted to various environments, and to predict the response of organisms to new or future conditions. The availability of large genomic and environmental data sets provides an opportunity to address those questions, and the R package LEA has been introduced to facilitate population and ecological genomic analyses in this context. By using latent factor models, the program computes ancestry coefficients from population genetic data, and performs genotype-environment association analyses with correction for unobserved confounding variables. In this study, we present new functionalities of LEA, which include imputation of missing genotypes, fast algorithms for latent factor mixed models using multivariate predictors for genotype-environment association studies, population differentiation tests for admixed or continuous populations, and estimation of genetic offset based on climate models. The new functionalities are implemented in version 3.0 and higher releases of the package. Using simulated and real data sets, our study provides evaluations and examples of applications, outlining important practical considerations when analyzing ecological genomic data in R.

International audience

International audience

In this thesis, we investigate resource allocation games, broadly defined as resource allocation problems involving interactions between competitive decision makers. We primarily focus on the Colonel Blotto game (CB game). In the CB game, two competitive players, each having a fixed budget, simultaneously distribute their resources toward n battlefields. Each player evaluates each battlefield with a certain value. In each battlefield, the player who has the higher allocation wins and gains the corresponding value. Each player's payoff is her aggregate gains from all the battlefields. First, we model several variants of the CB game and their extensions as one-shot complete-information games and analyze players' strategic behaviors. Our first main contribution is a class of approximate equilibria in these games for which we prove that the approximation error can be well-controlled. Second, we model resource allocation games with combinatorial structures as online learning problems to study situations involving sequential plays and incomplete information. We make a connection between these games and online shortest path problems. Our second main contribution is a set of novel regret-minimization algorithms for online shortest path problems under several feedback settings that provide significant improvements in regret guarantees and running time in comparison with existing solutions.; Dans cette thèse, nous étudions les jeux d'allocation de ressources généralement définis comme des problèmes d'allocation de ressources impliquant des interactions entre des décideurs compétitifs. Nous nous concentrons principalement sur le jeu Colonel Blotto (jeu CB). Dans ce jeu, deux joueurs compétitifs, chacun ayant un budget fixe, distribuent simultanément leurs ressources vers n champs de bataille. Chaque joueur évalue chaque champ de bataille avec une certaine valeur. Dans chaque champ de bataille, le joueur qui a l'allocation la plus élevée gagne la valeur correspondante. Le gain de chaque joueur est égal ses gains cumulés sur tous les champs de bataille. Nous modélisons plusieurs variantes et extensions du jeu CB comme jeux d'informations complètes à un coup. Notre première contribution est une classe d'équilibres approximatifs dans ces jeux tel que l'erreur d'approximation est bien contrôlée. Deuxièmement, nous modélisons les jeux d'allocation de ressources avec des structures combinatoires comme des problèmes d'apprentissage en ligne pour étudier des situations impliquant des jeux séquentiels et des informations incomplètes. Nous établissons une connexion entre ces jeux et les problèmes de chemin le plus court en ligne (OSP). Nous proposons nouveaux algorithmes d’OSP sous plusieurs paramètres de feedback qui améliorent des garanties de regret et du temps d'exécution.

Resource allocation problems are broadly defined as situations involving decisions on distributing a limited budget of resources in order to optimize an objective. In particular, many of them involve interactions between competitive decision-makers which can be well captured by game-theoretic models. In this thesis, we choose to investigate resource allocation games. We primarily focus on the Colonel Blotto game (CB game). In the CB game, two competitive players, each having a fixed budget of resources, simultaneously distribute their resources toward n battlefields. Each player evaluates each battlefield with a certain value. In each battlefield, the player who has the higher allocation wins and gains the corresponding value while the other loses and gains zero. Each player's payoff is her aggregate gains from all the battlefields. First, we model several prominent variants of the CB game and their extensions as one-shot complete-information games and analyze players' strategic behaviors. Our first main contribution is a class of approximate (Nash) equilibria in these games for which we prove that the approximation error can be well-controlled. Second, we model resource allocation games with combinatorial structures as online learning problems to study situations involving sequential plays and incomplete information. We make a connection between these games and online shortest path problems (OSP). Our second main contribution is a set of novel regret-minimization algorithms for generic instances of OSP under several restricted feedback settings that provide significant improvements in regret guarantees and running time in comparison with existing solutions., Les problèmes d'allocation des ressources sont définis comme les situations concernant les décisions sur la distribution d’un budget limité afin d’optimiser un objectif. Beaucoup d'entre eux impliquent des interactions entre des décideurs compétitifs ; ils peuvent être bien capturés par des modèles de théorie des jeux. Dans cette thèse, nous choisissons d'étudier les jeux d'allocation de ressources. Nous nous concentrons principalement sur le jeu de Colonel Blotto (CB). Dans le jeu CB, deux joueurs compétitifs, chacun ayant un budget fixe, distribuent simultanément leurs ressources vers n champs de bataille. Chaque joueur évalue chaque champ de bataille avec une certaine valeur. Dans chaque champ de bataille, le joueur qui a l'allocation la plus élevée gagne la valeur correspondante tandis que l'autre obtient zéro. Le gain de chaque joueur est à ses gains cumulés sur tous les champs de bataille. Tout d'abord, nous modélisons plusieurs variantes et extensions du jeu CB comme jeux d'informations complètes à un coup. Notre première contribution est une classe d'équilibres approximatifs dans ces jeux et nous prouvons que l'erreur d'approximation est bien contrôlée. Deuxièmement, nous modélisons les jeux d'allocation de ressources avec des structures combinatoires comme des problèmes d'apprentissage en ligne pour étudier des situations impliquant des jeux séquentiels et des informations incomplètes. Nous établissons une connexion entre ces jeux et les problèmes de chemin le plus court en ligne (OSP). Notre deuxième contribution est un ensemble de nouveaux algorithmes d’OSP sous plusieurs paramètres de feedback qui améliorent des garanties de regret et du temps d'exécution.

Directed fuzzing focuses on automatically testing specific parts of the code by taking advantage of additional information such as (partial) bug stack trace, patches or risky operations. Key applications include bug reproduction, patch testing and static analysis report verification. Although directed fuzzing has received a lot of attention recently, hard-to-detect vulnerabilities such as Use-After-Free (UAF) are still not well addressed, especially at the binary level. We propose UAFuzz, the first (binary-level) directed greybox fuzzer dedicated to UAF bugs. The technique features a fuzzing engine tailored to UAF specifics, a lightweight code instrumentation and an efficient bug triage step. Experimental evaluation for bug reproduction on real cases demonstrates that UAFuzz significantly outperforms state-of-the-art directed fuzzers in terms of fault detection rate, time to exposure and bug triaging. UAFuzz has also been proven effective in patch testing, leading to the discovery of 30 new bugs (7 CVEs) in programs such as Perl, GPAC and GNU Patch. Finally, we provide to the community a large fuzzing benchmark dedicated to UAF, built on both real codes and real bugs., The long version of paper appeared in the 23rd International Symposium on Research in Attacks, Intrusions and Defenses (RAID 2020)

We present an efficient approach to retrieve behavioural models from reactive software systems in the form of Finite State Machines by testing them. The system is accessed in black box mode; thus, no source or binary code is needed. The novelty of the approach is that it does not require to reset the system between tests (queries) and does not require any knowledge of the system apart from its input domain. Experiments have shown that it can scale up to systems that may have thousands of states.

International audience; This work tackles the reachability problem of uncertain discrete-time linear systems controlled by estimated state feedback regulators. First an augmented system is considered in order to include the effect of the estimation error on the behavior of the closed loop system. Then, based on interval analysis, an interval predictor is proposed to compute tight trajectory tubes that contain in a guaranteed way the actual state trajectory of the augmented system. Moreover, under the standard controllability and observability assumptions of linear systems, the convergence of the width of these trajectory tubes is shown.

International audience; Les methodes expeŕimentales dans la recherche en management international RESUME Les méthodes expérimentales sont depuis longtemps employées en Sciences naturelles afin de tester une hypothèse en contrôlant les conditions d'expérimentation. Ces méthodes ont été reprises en Sciences économiques et de gestion ces dernières décennies, et sont devenues une référence dans certaines sous-disciplines comme le marketing. Or, les expérimentations restent rares dans la recherche en management international. Dans ce chapitre, nous présentons les difficultés, mais aussi les avantages des méthodologies expérimentales dans cette sous-discipline à travers l'exemple du projet InterCCom, portant sur la compétence interculturelle. Grâce au développement d'un jeu sérieux informatisé dans le cadre de ce projet, les comportements des membres d'une équipe internationale virtuelle sont mesurés, ce qui permettra de réponse à des problématiques de recherche variées.

International audience; Les methodes expeŕimentales dans la recherche en management international RESUME Les méthodes expérimentales sont depuis longtemps employées en Sciences naturelles afin de tester une hypothèse en contrôlant les conditions d'expérimentation. Ces méthodes ont été reprises en Sciences économiques et de gestion ces dernières décennies, et sont devenues une référence dans certaines sous-disciplines comme le marketing. Or, les expérimentations restent rares dans la recherche en management international. Dans ce chapitre, nous présentons les difficultés, mais aussi les avantages des méthodologies expérimentales dans cette sous-discipline à travers l'exemple du projet InterCCom, portant sur la compétence interculturelle. Grâce au développement d'un jeu sérieux informatisé dans le cadre de ce projet, les comportements des membres d'une équipe internationale virtuelle sont mesurés, ce qui permettra de réponse à des problématiques de recherche variées.

International audience; The water-processable microfibrillated cellulose (MFC) and sodium lignosulfonate (NaLS) system was used for the first time as carbon precursor. After drying and one-step pyrolysis up to 800 °C, the resulting bio carbonaceous char can acquire considerable electrical conductivity and porosity, making it a promising electrode material in energy storage devices. The present work characterizes MFC and NaLS separately as well as in blends in terms of morphology (SEM), structure (FTIR) and thermal stability (TG–MS). DTG patterns illustrate that MFC degrades in a narrow temperature range but at high reaction rates, whereas NaLS reacts more steadily in a larger temperature range. The thermal degradation rate of a MFC/NaLS blend differs from an additive weighted thermal degradation rate of each component: MFC and NaLS. The presence of sodium in the blends induced catalytic effects and made MFC decomposition to shift toward lower temperatures. This observation could be kinetically interpreted as a decrease of the MFC thermal degradation activation energies in the composites. The kinetics of MFC/NaLS blends is investigated by both model-free and mode-based methods, with the latter providing more appropriate Arrhenius parameters that allow reproducing the experimental curves

Efficient detection and instrumentation of function calls is fundamental for a variety of dynamic analysis techniques, including dynamic callgraph construction, control-flow integrity, and automatic vulnerability discovery. A common way of detecting calls at the machine code level is to look for CALL instructions. However, optimizing compilers frequently implement function tail calls with JMP instructions instead, and distinguishing an intra-procedural jump from a JMP-based function call is not straightforward. Despite the importance of making this distinction, prior research has not produced a reliable solution. In this paper, we address the problem of dynamic function call detection in real-time. We propose a heuristic-based approach named iCi to efficiently and automatically instrument calls, including conventional CALLs and JMP-based calls, at runtime. iCi does not rely on source code, debug information, symbol tables or static analysis. We show that iCi achieves an f-score of 0.95 in the worst case, regardless of optimization level. We open-source our implementation as well as the oracle we used for our evaluation.

Active inference algorithms that are used to extract behavioural models of software systems usually assume that the System Under Inference (SUI) can be reset. Two approaches have been proposed to infer systems that cannot be reset. Rivest and Schapire proposed an adaptation of the \(L^*\) algorithm that relies on having a homing sequence for the SUI. We detail here another approach that is based on characterization sequences. More precisely, we assume classical testing hypotheses, namely that we are given a bound n on the number of states and a set W of characterizing sequences to distinguish states. Contrary to \(L^*\), it does not require an external oracle to decide on equivalence. The length of the test sequence is polynomial in n and the exponent depends on the cardinality |W| of the characterization set. For systems where resetting is impossible or expensive, this approach can be a viable alternative to classical learning methods.

Machine learning in the form of inference of state machine models has gained popularity in model-based testing as a means of retrieving models from software systems. By combining an old idea from machine inference with methods from automata testing in a heuristic approach, we propose a new promising direction for inferring black box systems that cannot be reset. Preliminary experiments show that this heuristic approach scales up well and outperforms more systematic approaches.

The paper focuses on the problems of passive and active FSM inference as well as checking sequence generation. We consider the setting where an FSM cannot be reset so that its inference is constrained to a single trace either given a priori in a passive inference scenario or to be constructed in an active inference scenario or aiming at obtaining checking sequence for a given FSM. In each of the last two cases, the expected result is a trace representing a checking sequence for an inferred machine, if it was not given. We demonstrate that this can be achieved by a repetitive use of a procedure that infers an FSM from a given trace (identifying a minimal machine consistent with a trace) avoiding equivalent conjectures. We thus show that FSM inference and checking sequence construction are two sides of the same coin. Following an existing approach of constructing conjectures by SAT solving, we elaborate first such a procedure and then based on it the methods for obtaining checking sequence for a given FSM and inferring a machine from a black box. The novelty of our approach is that it does not use any state identification facilities. We demonstrate that the proposed approach can also be constrained to find a solution in a subset of FSMs represented by a nondeterministic mutation machine. Experiments with a prototype implementation of the developed approach using an existing SAT solver indicate that it scales for FSMs with up to a dozen of states and requires relatively short sequences to identify a black box machine.

International audience; —When designing new solutions for Low Power Wide Area (LPWA) networks, coexistence and integration into existing 4G frameworks should be considered to ease the deployment procedure and reduce costs. In a previous work, Turbo-Frequency Shift Keying (FSK) was proposed as a potential physical layer for LPWA networks. We propose to study the system in the Orthogonal Frequency Division Multiplexing (OFDM) framework , currently used by existing cellular networks and also considered for the recently standardized Narrow-Band IoT. The new architectures of the Turbo-FSK transmitter and receiver are presented. Simulations are run under various conditions including Rayleigh fading channel with mobility, demonstrating the Turbo-FSK scheme to achieve performance close to Turbo-Coded OFDM. The scheme also exhibits a constant envelope, which may multiply the Power Amplifier (PA) efficiency by up to a factor of 2 and makes the solution a promising candidate for 5G LPWA transmission.