Search

International audience; We consider the problem of learning time-consuming functions defined over unordered sets of vectors. Such functions arise frequently, in particular in the context of networks of devices whose number is not fixed and that interact with each other. A working example is the modeling of a wind farm. Unordered sets of vectors are a mix of integer and continuous input variables suitable for functions that are permutation-invariant. The time-consuming aspect of the functions is, classically, treated by approximating them with a Gaussian process.This study addresses the problem of defining valid and efficient covariance kernels over cloudsof points in the context of Gaussian process surrogate modeling. We review methods for defining such kernels. These kernels are compared on a set of analyticalfunctions inspired from different engineering problems, such as the design of experiments andthe modeling of wind farms production. The extrapolation properties of the kernels are testedon geometrically transformed clouds.We show that modeling 2D clouds of points as supports of discrete uniform distributions should be preferred to a Gaussian representation of the clouds. A detailed investigation of the good performance of MMD-based kernels illustrates how they adapt their hyperparameters to the geometrical properties of the studied functions.

International audience; We consider the task of estimating functions from a restricted number of observations where the inputs are in the form of varying-size sets of vectors. A classical method in this expensive functions context is to approximate the true expensive function with a Gaussian process that relies on semi-definite positive kernels. Varying-size sets of vectors have some peculiar characteristics : the number of vectors is not fixed and the vectors are not ordered. These properties prevent the use of classical, fixed dimension, kernels. Sets of vectors can be represented as probability distributions or as vectors composed of wellchosen characteristic features. The distribution associated with each set of vectors can be either a (non-informative) discrete uniform distribution supported by the points of the set, or an empirical Gaussian distribution characterized by a mean and a covariance matrix estimated from the vectors of the set as in [5]. With this representation, we can define semi-definite positive kernels between distributions or vectors. The kernels can be classified into explicit ones, with known feature map and scalar product, and implicit kernels where the feature maps and the scalar product come mixed together. For implicit kernels, we restrict ourselves to distance substitution kernels with an exponential form. It is known that this approach for constructing kernels requires the substitute to be conditionally semi-definite negative, see [1]. In this context, it is necessary and sufficient to have an Hilbertian metric (isometric to a L 2 norm) [4]. The distances between distributions verifying the latter condition which we consider are the sliced Wasserstein distance [3], the maximum mean discrepancy, the approximated Gaussian Wasserstein (see [2]), and the Euclidean distance for vectors of features. For explicit kernels, we investigate the probability product kernel between Gaussian distributions [5] and the scalar product of embeddings of discrete uniform laws. The prediction performance of theses kernels are compared on three types of analytical functions. The first ones mimic wind-farms productions for different winds, the second function is the inertia of a set of vectors and the third is the maximin function, well-known in the design of experiments. We first analyze the predictions over sets of vectors drawn uniformly over the same rectangular search space. Afterwards, we investigate the predictions over sets of vectors modified through geometrical transformations such as dilatation, rotation and translation. We show that some geometrical properties of the wind-farm functions can be better learned by MMD-based kernels than others. As an example, the MMD-based kernel prediction together with the true function is represented in Figure 1.

International audience; Deep neural networks have the capacity to generate textual content which is increasingly difficult to distinguish from that produced by humans. Such content can be used in disinformation campaigns and its detrimental effects are amplified if it spreads on social networks. Here, we study the automatic detection of bot-generated Twitter messages. This task is difficult due to combination between the strong performance of recent deep language models and the limited length of tweets. In this study, we propose a challenging definition of the problem by making no assumption regarding the bot account, its network or the method used to generate the text. We devise two approaches for bot detection based on pretrained language models and create a new dataset of generated tweets to improve the performance of our classifier on recent text generation algorithms. The obtained results show that the generalization capabilities of the proposed classifier heavily depends on the dataset used to trained the model. Interestingly, the two automatic dataset augmentation proposed here show promising results. Their introduction leads to consistent performance gains compared to the use of the original dataset alone.

International audience

Objective:To describe the clinical aspects of urethral strictures managed internal endoscopic urethrotomy (IEU) and to evaluate the outcomes and factors on success rate. Patients and methods:This is a retrospective, descriptive, analytical and monocentric 2-year study of patient records with an IEU indicated for urethral stricture. The parameters studied were: age, history of endo-urethral maneuvers, clinical symptomatology,etiologies, strictures characteristics and outcomes. The outcomes were appreciated by the quality of the voiding jet and the outcomes of the control rétrograde urethrocystography.. Results:Sixty patients were included. The mean age was 50±19.3 years old. An history endo-urethral was found in 38.3% of patients. The mean of consultation time was 20.7±30.7 months. Dysuria was the most common reason for consultation. The most common etiologies were those of infectious and iatrogenic. The bulbar seat was predominant. The mean of length was 0.9±0.5cm. After the mean delay of 14.5 ± 6.9 months, the overall success rate was 60%. Per and postoperative complications were dominated by retention of urine. The outcomes showed that there was no significant correlation between age, previous treatment, etiology, seat, length, number of strictures, and outcome. Conclusion:The IEU gives a good results if the indication is well posed. Seat, length, single or multiple character, etiology and history of endo-urethral maneuvers do not affect the outcomes.&nbsp