HAL Paris Dauphine-PSL
Not a member yet
21932 research outputs found
Sort by
Livestock Farming Trapped in Antimicrobial Reduction: How to Put Just Transition at the Heart of AMR Policies?
No full textInternational audienceThis article examines how antimicrobial resistance (AMR) policies designed for livestock farming in Europe—and particularly France—have been limited by narrowly defined reduction targets. Although these policies have significantly decreased antibiotic use, they have also upheld a productivist agricultural model that continues to threaten human and animal health and the environment. By aligning AMR mitigation efforts with a biosecurity paradigm that is highly compatible with industrial livestock systems, the reduction in antibiotic use has not yielded all the anticipated benefits. Based on this assessment, we propose three avenues for fostering a just transition relevant to all sectors involved in AMR governance: transcending the dominant “One Global Health” paradigm; shifting power from institutional stakeholders to the public affected by AMR; and reimagining post-antibiotic futures that extend beyond prevailing dystopian narratives
SigGate: Enhancing Recurrent Neural Networks with Signature-Based Gating Mechanisms
No full textIn this paper, we propose a novel approach that enhances recurrent neural networks (RNNs) by incorporating path signatures into their gating mechanisms. Our method modifies both Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) architectures by replacing their forget and reset gates, respectively, with learnable path signatures. These signatures, which capture the geometric features of the entire path history, provide a richer context for controlling information flow through the network's memory. This modification allows the networks to make memory decisions based on the full historical context rather than just the current input and state. Through experimental studies, we demonstrate that our Signature-LSTM (SigLSTM) and Signature-GRU (SigGRU) models outperform their traditional counterparts across various sequential learning tasks. By leveraging path signatures in recurrent architectures, this method offers new opportunities to enhance performance in time series analysis and forecasting applications
Scalable and Adaptive Variational Bayes Methods for Hawkes Processes
No full textInternational audienceHawkes processes are often applied to model dependence and interaction phenomena in multivariate event data sets, such as neuronal spike trains, social interactions, and financial transactions. In the nonparametric setting, learning the temporal dependence structure of Hawkes processes is generally a computationally expensive task, all the more with Bayesian estimation methods. In particular, for multivariate nonlinear Hawkes processes, Monte-Carlo Markov Chain (MCMC) methods used to sample from the posterior distribution do not scale well to the dimension of the process. Recently, efficient algorithms targeting a mean-field variational approximation of the posterior distribution have been proposed, however, these methods do not allow to perform model selection on the graph of interactions of the Hawkes model. In this work, we propose a novel adaptive Bayesian variational method that performs model selection and can estimate a sparse graphical parameter. For the popular sigmoid Hawkes processes, we design a parallel algorithm which is scalable to high-dimensional point processes and large sequences of events. Furthermore, we unify existing variational Bayes approaches under a general nonparametric inference framework, and analyse the asymptotic properties of these methods under easily verifiable conditions on the prior, the variational class, and the nonlinear model. Finally, through an extensive set of numerical simulations, we demonstrate that our method is able to adapt to the dimensionality of the parameter of the Hawkes process, and is partially robust to certain types of model misspecification
La résolution des conflits entre associés par l’exclusion ou le retrait
No full textNational audienceLa communication envisage le retrait et l'exclusion comme modalités possibles de résolution d'un conflit entre associés, dans les différentes sociétés d'exercice des professions libérales réglementées (SCP, SEL et SPE)
Recurrent Neural Networks for Dynamic VWAP Execution: Adaptive Trading Strategies with Temporal Kolmogorov-Arnold Networks
No full textThe execution of Volume Weighted Average Price (VWAP) orders remains a critical challenge in modern financial markets, particularly as trading volumes and market complexity continue to increase. In my previous work [1], I introduced a novel deep learning approach that demonstrated significant improvements over traditional VWAP execution methods by directly optimizing the execution problem rather than relying on volume curve predictions. However, that model was static because it employed the fully linear approach described in [2], which is not designed for dynamic adjustment. This paper extends that foundation by developing a dynamic neural VWAP framework that adapts to evolving market conditions in real time. We introduce two key innovations: first, the integration of recurrent neural networks to capture complex temporal dependencies in market dynamics, and second, a sophisticated dynamic adjustment mechanism that continuously optimizes execution decisions based on market feedback. The empirical analysis, conducted across five major cryptocurrency markets, demonstrates that this dynamic approach achieves substantial improvements over both traditional methods and our previous static implementation, with execution performance gains of 10-15% in liquid markets and consistent outperformance across varying conditions. These results suggest that adaptive neural architectures can effectively address the challenges of modern VWAP execution while maintaining computational efficiency suitable for practical deployment
Managers' experience of boredom: a reading through their strategic roles
No full textInternational audienceStrategic management has shown little or no interest in the problem of boredom in companies, preferring to leave it to human resources management and psychology, perhaps on the pretext that boredom is a characteristic of certain types of operational work. Boredom also concerns strategy when we are interested in its social dimension and when it affects certain categories of people, such as middle managers, not studied in the literature on boredom. The social approach then focuses on boredom as a subjective experience, which cannot be approached outside the social and relational environment in which it occurs: the organisation, its strategy, its culture or, more broadly, the contemporary world. It becomes important to study workers' experience of boredom as a socially situated phenomenon, rooted in specific organisational work practices. Such an approach highlights that boredom can not only be the translation of a job well done but also be a source of innovation, and engage bored individuals in a process of creating new forms of organisation. In this research we are interested in the social dimension of middle managers' boredom by seeking to understand their experience of boredom in the context of the exercise (or not) of their strategic roles within organisation. To do this, we conducted 30 interviews with middle managers in the banking sector who are experiencing, or have experienced, boredom. We coded the interviews using a thematic approach based on Floyd and Wooldridge's typology of the strategic roles of middle managers. Our results allow us to conclude firstly that middle managers can experience boredom, which they link to the evolution of their profession as managers in a sector undergoing rapid change. Second, this experience can be understood by their ability to exercise some of their strategic roles, particularly in the context of a strategy deliberately imposed by top management, but above all by the fact that they feel increasingly impeded from exercising strategic influence, particularly in the context of a more emerging strategy. Then they question their strategic positioning within their organisation, feeling that they are expected to perform mainly operational tasks. We thus contribute to the literature on boredom by highlighting the importance of its social dimension for middle managers. We also contribute to management strategic literature discussing ways out of boredom by suggesting that middle managers reinvent their strategic role, particularly in relation to their superiors, but also with their fellow managers
The multiple steiner TSP with cyclic order on terminals: valid inequalities and polyhedra
No full textInternational audienceThis paper deals with a variant of the Traveling Salesman Problem (TSP), called the Multiple Steiner TSP with Order Constraints (MSTSPOC). Consider an undirected graph with nonnegative weights on the edges, and a set of salesmen such that with each salesman is associated a set of ordered terminals. The MSTSPOC consists in finding a minimum-weight subgraph containing for each salesman a tour going in order through its terminals. We study the polytope associated with the Integer Linear Programming (ILP) formulation proposed in Borne et al. (2013). We characterize when the basic inequalities define facets. We also describe new valid inequalities along with necessary conditions and sufficient conditions for these inequalities to be facet defining. Further families of valid inequalities, coming from closely related problems, are also discussed. The theoretical results presented in this paper are computationally tested in a companion paper (Taktak 2024)
Generalized Bayesian deep reinforcement learning
No full textInternational audienceBayesian reinforcement learning (BRL) is a method that merges principles from Bayesian statistics and reinforcement learning to make optimal decisions in uncertain environments. As a model-based RL method, it has two key components: (1) inferring the posterior distribution of the model for the data-generating process (DGP) and (2) policy learning using the learned posterior. We propose to model the dynamics of the unknown environment through deep generative models, assuming Markov dependence. In the absence of likelihood functions for these models, we train them by learning a generalized predictive-sequential (or prequential) scoring rule (SR) posterior. We used sequential Monte Carlo (SMC) samplers to draw samples from this generalized Bayesian posterior distribution. In conjunction, to achieve scalability in the high-dimensional parameter space of the neural networks, we use the gradient-based Markov kernels within SMC. To justify the use of the prequential scoring rule posterior, we prove a Bernstein-von Mises-type theorem. For policy learning, we propose expected Thompson sampling (ETS) to learn the optimal policy by maximising the expected value function with respect to the posterior distribution. This improves upon traditional Thompson sampling (TS) and its extensions, which utilize only one sample drawn from the posterior distribution. This improvement is studied both theoretically and using simulation studies, assuming a discrete action space. Finally, we successfully extended our setup for a challenging problem with a continuous action space without theoretical guarantees
