HAL Paris Dauphine-PSL
Not a member yet
21932 research outputs found
Sort by
Lessons in Legitimacy and the EU Commitment to Environmental Democracy: Notes from MOP8 of the Aarhus Convention
No full texthttps://2035legitimacy.fi/lessons-in-legitimacy-and-the-eu-commitment-to-environmental-democracy-notes-from-mop8-of-the-aarhus-convention/Last week, as the world's attention focused on the COP30 in Belin, another international meeting gathered in the far cooler setting of Lake Geneva. The participants at this meeting were also preoccupied with how the world responds to the triple planetary crisis, and with the fate of multilateralism and international institutions. And at this meeting too, the delicate balance of power between citizens and states was askew.</div
CoHiRF: A Scalable and Interpretable Clustering Framework for High-Dimensional Data
No full textClustering high-dimensional data poses significant challenges due to the curse of dimensionality, scalability issues, and the presence of noisy and irrelevant features. We propose Consensus Hierarchical Random Feature (CoHiRF), a novel clustering method designed to address these challenges effectively. CoHiRF leverages random feature selection to mitigate noise and dimensionality effects, repeatedly applies K-Means clustering in reduced feature spaces, and combines results through a unanimous consensus criterion. This iterative approach constructs a cluster assignment matrix, where each row records the cluster assignments of a sample across repetitions, enabling the identification of stable clusters by comparing identical rows. Clusters are organized hierarchically, enabling the interpretation of the hierarchy to gain insights into the dataset. CoHiRF is computationally efficient with a running time comparable to K-Means, scalable to massive datasets, and exhibits robust performance against state-of-the-art methods such as SC-SRGF, HDBSCAN, and OPTICS. Experimental results on synthetic and real-world datasets confirm the method's ability to reveal meaningful patterns while maintaining scalability, making it a powerful tool for high-dimensional data analysis
Speaker Group Encoding in Self-supervised Speech Recognition Models
No full textInternational audienceWe investigate what self-supervised speech recognition models (S3Ms) learn about speaker groups (SGs). We examine several states of S3Ms: pretrained, finetuned on speaker identification (SID), as well as finetuned automatic speech recognition (ASR) using fairness enhancing algorithms. We find that S3Ms encode information about several speaker group categories (SGCs), including their gender, age, dialect, ethnicity, and whether they are a native speaker. We find that finetuning for SID amplifies certain SGCs, namely those whose variance is more phonetic in nature. Meanwhile, finetuning for ASR discards phonetically variant speaker group information (SGI) but retains semantically variant SGI. We find that ASR algorithms designed for fairness improvement change to what extent SGI is encoded in S3Ms; however, this is primarily true for phonetically variant SGCs, and less true for semantically variant SGCs. We discuss how SGI is encoded by each layer, and identify subdimensions of embeddings responsible for encoding different SGCs. Finally, we discuss how our findings could be beneficial in designing fairer ASR algorithms
Minimizing Rosenthal's Potential in Monotone Congestion Games
No full textInternational audienceCongestion games are attractive because they can model many concrete situations where some competing entities interact through the use of some shared resources, and also because they always admit pure Nash equilibria which correspond to the local minima of a potential function. We explore the problem of computing a state of minimum potential in this setting. Using the maximum number of resources that a player can use at a time, and the possible symmetry in the players' strategy spaces, we settle the complexity of the problem for instances having monotone (i.e., either non-decreasing or non-increasing) latency functions on their resources. The picture, delineating polynomial and NP-hard cases, is complemented with tight approximation algorithms.</div
Heterogeneous Trade Effects of Pre-Shipment Inspections
No full textInternational audienceWe analyze the trade impact of pre-shipment inspections (PSI)—a practice under which imports need to undergo a third-party review process before shipment, and whose utilization has been limited by the WTO Trade Facilitation Agreement. We show that PSI requirements had a negative impact on imports, and were most harmful for trade in differentiated manufacturing products (administrative trade cost channel). In contrast, PSI were facilitating trade in products subject to conformity assessment procedures related to sanitary and phytosanitary measures (information channel). Counterfactual analysis suggests that the reduction in administrative costs outweighed the provision of information, with the removal of PSI leading to a slight increase in developing countries’ imports. The removal of PSI could also induce a cost: the relaxation of controls on custom misinvoicing. We show that PSI had a limited effect on trade misinvoicing at the intensive margin, and on lost exports at the extensive margin
Drift Estimation for Diffusion Processes Using Neural Networks Based on Discretely Observed Independent Paths
No full textThis paper addresses the nonparametric estimation of the drift function over a compact domain for a time-homogeneous diffusion process, based on high-frequency discrete observations from N independent trajectories. We propose a neural network-based estimator and derive a non-asymptotic convergence rate, decomposed into a training error, an approximation error, and a diffusion-related term scaling as log N /N . For compositional drift functions, we establish an explicit rate. In the numerical experiments, we consider a drift function with local fluctuations generated by a double-layer compositional structure featuring local oscillations, and show that the empirical convergence rate becomes independent of the input dimension d. Compared to the B-spline method, the neural network estimator achieves better convergence rates and more effectively captures local features, particularly in higher-dimensional settings
Velocity Trapping in the Lifted Totally Asymmetric Simple Exclusion Process and the True Self-Avoiding Random Walk
No full textInternational audienceWe discuss nonreversible Markov-chain Monte Carlo algorithms that, for particle systems, rigorously sample the positional Boltzmann distribution and that have faster than physical dynamics. These algorithms all feature a nonthermal velocity distribution. They are exemplified by the lifted totally asymmetric simple exclusion process (lifted TASEP), a one-dimensional lattice reduction of event-chain Monte Carlo. We analyze its dynamics in terms of a velocity trapping that arises from correlations between the local density and the particle velocities. This allows us to formulate a conjecture for its out-of-equilibrium mixing timescale, and to rationalize its equilibrium superdiffusive timescale. Both scales are faster than for the (unlifted) TASEP. They are further justified by our analysis of the lifted TASEP in terms of many-particle realizations of true self-avoiding random walks. We discuss velocity trapping beyond the case of one-dimensional lattice models and in more than one physical dimensions. Possible applications beyond physics are pointed out
Les droits de la défense des agents publics : esquisse d'un état des lieux nuancé de la dé-procéduralisation
No full textLes droits de la défense des agents publics permettent à ces derniers de faire valoir leur point de vue avant qu'une sanction disciplinaire ou une mesure prise en considération de leur personne ne soit prise à leur encontre. Assouplissement au caractère unilatéral de la décision exécutoire, ces droits ont été opposés, dès la loi du 22 avril 1905, à l'administration et se sont développés continûment sous l'action du législateur et des juges. Aujourd'hui, ces droits ne sont pas remis en cause par les différentes jurisprudences, ni même par la suppression de certaines commissions, sans préjudice des effets de la nouvelle procédure de rupture conventionnelle dans la fonction publique, qui les écartent de plano
Nonparametric Bayesian intensity estimation for covariate-driven inhomogeneous point processes
No full textInternational audienceThis work studies nonparametric Bayesian estimation of the intensity function of an inhomogeneous Poisson pointprocess in the important case where the intensity depends on covariates, based on the observation of a singlerealisation of the point pattern over a large area. It is shown how the presence of covariates allows to borrow infor-mation from far away locations in the observation window, enabling consistent inference in the growing domainasymptotics. In particular, optimal posterior contraction rates under both global and point-wise loss functions arederived. The rates in global loss are obtained under conditions on the prior distribution resembling those in thewell established theory of Bayesian nonparametrics, combined with concentration inequalities for functionals ofstationary processes to control certain random covariate-dependent loss functions appearing in the analysis. Thelocal rates are derived with an ad-hoc study that builds on recent advances in the theory of Pólya tree priors, ex-tended to the present multivariate setting with a novel construction that makes use of the random geometry inducedby the covariates