HAL Paris Dauphine-PSL
Not a member yet
21932 research outputs found
Sort by
Birkhoff attractors for dissipative symplectic billiards
No full textThe aim of the present paper is to propose and study a dissipative variant of symplectic billiards within planar strictly convex domains. The associated billiard map is dissipative, thus it admits a compact invariant set, the so-called Birkhoff attractor. Its complexity depends on the rate of the dissipation as well as on the geometry of the billiard table. We prove that (a) for strong dissipation, the Birkhoff attractor is a normally contracted graph over the zero section; (b) for mild dissipation, the Birkhoff attractor within a centrally symmetric domain is an indecomposable continuum whose restricted dynamics has positive topological entropy. We compare these results with the case of dissipative Birkhoff billiards, studied in Bernardi, Florio and Leguil's paper
High order universal portfolios
No full textInternational audienceThe Cover universal portfolio has many interesting theoretical and numerical properties and was investigated for a long time. Building on it, we explore what happens when we add this UP to the market as a new synthetic asset and construct by recurrence higher order UPs. We investigate some important theoretical properties of the high order UPs and show in particular that they are indeed different from the Cover UP and are capable to break the time permutation invariance. We show that under some perturbation regime the second order UP has better Sharp ratio than the standard UP and briefly investigate arbitrage opportunities thus created. Numerical experiences on a benchmark from the literature confirm that high order UPs improve Cover’s UP performances
A comparative analysis of risk prevention policy tools and governance structures in Normandy (France) and Victoria (Australia): Assessing policies for high-risk sites
No full textInternational audienceHigh-risk industrial sites are classified as the most dangerous sites for their risk of inducing damage in case of an accident, even if it is rare. Risk policies are one way to prevent damage while balancing other trade-offs. How can public authorities assess "risk policy systems" of high-risk sites in specific administrative regimes for an effective decision process taking into consideration different jurisdictions? This paper investigates the Normandy and Victorian risk policy system in France and Australia, respectively, based on qualitative approaches. By investigating two cases and fostering conceptual thinking, two assessment criteria are highlighted by different stakeholders to determine an effective risk policy. For instance, beyond the common focus on the efficiency and efficacy of risk policies, findings revealed the importance of "dialogue, participation, and cooperation" between actors and "adaptivity, flexibility, relevance, and coherence" of different tools and legal texts. Limitations and further work include testing and iterating on other assessment criteria to propose an adapted assessment framework for high-risk policies
La place des dépenses fiscales et socio-fiscales dans le système des prélèvements et dépenses publics en France depuis 1979 : une base de données inédite
No full textInternational audienc
Pseudo-MDPs: A Novel Framework for Efficiently Optimizing Last Revealer Seed Manipulations in Blockchains
No full textThis study tackles the computational challenges of solving Markov Decision Processes (MDPs) for a restricted class of problems. It is motivated by the Last Revealer Attack (LRA), which undermines fairness in some Proof-of-Stake (PoS) blockchains such as Ethereum ($400B market capitalization). We introduce pseudo-MDPs (pMDPs) a framework that naturally models such problems and propose two distinct problem reductions to standard MDPs. One problem reduction provides a novel, counter-intuitive perspective, and combining the two problem reductions enables significant improvements in dynamic programming algorithms such as value iteration. In the case of the LRA which size is parameterized by κ (in Ethereum's case κ = 32), we reduce the computational complexity from O(2^κ κ^2^(κ+2)) to O(κ^4) (per iteration). This solution also provide the usual benefits from Dynamic Programming solutions: exponentially fast convergence toward the optimal solution is guaranteed. The dual perspective also simplifies policy extraction, making the approach well-suited for resource-constrained agents who can operate with very limited memory and computation once the problem has been solved. Furthermore, we generalize those results to a broader class of MDPs, enhancing their applicability. The framework is validated through two case studies: a fictional card game and the LRA on the Ethereum random seed consensus protocol. These applications demonstrate the framework's ability to solve large-scale problems effectively while offering actionable insights into optimal strategies. This work advances the study of MDPs and contributes to understanding security vulnerabilities in blockchain systems
Minimal Variance Allocation of Rights and Applications to Blockchain Consensus
No full textRandomness plays a critical role in distributed systems and blockchain technology, facilitating tasks such as load balancing, leader election, and fault tolerance, which enhance system scalability and resilience. However, while randomness is fundamental to certain systems, it can also pose liabilities, including unpredictability and security risks. In this paper, we focus on the allocation of rights, a shared random process in many blockchains, particularly in the allocation of block proposal or validation rights. We propose two new allocations protocol to reduce the randomness of this process and discuss the challenges of achieving minimal variance in this allocation. Beyond improving the allocation protocol, they offer an alternative approach to secure random seed generation by reducing its criticality. In fact, this serves as a complementary solution to existing measures that focus on preventing random seed manipulation. In addition, we evaluate the effectiveness of our proposal and analyze potential countermeasures that attackers might employ. Our approach improves the resilience and security of blockchain networks, addressing concerns associated with randomness in consensus protocols
LLMs vs. econometric models for nowcasting GDP growth: A practitioner's view
No full textThis paper evaluates the performance of Large Language Models (LLMs) in nowcasting French GDP growth, comparing them with the econometric models currently used by the Banque de France. Using only prompt-based queries without external data or fine-tuning, the study assesses whether general-purpose LLMs such as ChatGPT, Gemini, and Claude can serve as effective forecasting tools. While econometric models consistently outperform LLMs during normal periods, the latter show a notable advantage in capturing exceptional events such as the COVID-19 pandemic. The paper also examines the sensitivity of LLM forecasts to prompt language, design, and model version, and introduces a confidence index and recession probability derived from LLM responses. There is no strong evidence of information leakage, and robustness checks confirm the findings across various model versions and temperature values. A fair insample comparison reinforces the relative strength of econometric models in normal conditions. Overall, the results suggest that while standard LLMs are not yet ready to replace traditional models in routine forecasting, they can provide complementary insights, particularly in periods of structural change or heightened uncertainty. These results apply to the most popular LLMs without external data and fine-tuning
Le travail journalistique entre priorisation de l’information et réaction aux algorithmes de recommandation des réseaux sociaux
No full textInternational audienceThis paper sheds light on the consequences of social network recommendation algorithms on journalistic work. It focuses particularly on the experience of French journalists through an investigation begun during the Yellow Vests protests—a social movement that benefited for its structuring from the modification of Facebook’s News Feed algorithm that occurred in 2018, and which expressed in a ritualized way a violence towards journalists pushing the latter to question some of their practices. Our work is based on more than fifty interviews with editorial directors, editors-in-chief, reporters and fact-checkers working in major national general-interest newsrooms (mainly Le Monde and Libération). Some of those interviewed were also founders of pure players (Les Jours, Rue89, Slate Fr, Loopsider, Orient XXI...). In addition, other respondents are heads (of training) of journalism schools (Sciences Po, IPJ Dauphine, ESJ Lille) or belong to professional associations (Reporters Without Borders, Prenons la Une). After a brief review of the inner workings of recommendation algorithms and their impact, we question the way in which these algorithms undermine the work of producing and prioritizing information. We also show how these algorithms reveal and reinforce socially situated journalistic practices and trends.Cet article propose de mettre en lumière les conséquences des algorithmes de recommandation des réseaux sociaux sur le travail journalistique. Il s’intéresse en particulier à l’expérience de journalistes français à travers une enquête commencée lors des protestations des Gilets jaunes, mouvement social ayant bénéficié pour sa structuration de la modification de l’algorithme du fil d’actualité de Facebook survenue en 2018, et exprimé de manière ritualisée une violence à l’égard des journalistes poussant ces derniers à s’interroger sur certaines de leurs pratiques. Notre travail se fonde notamment sur l’exploitation d’une cinquantaine d’entretiens avec des directeur·rice·s de rédactions, rédacteur·rice·s en chef, journalistes reporters ou fact-checkers, évoluant au sein de grandes rédactions nationales d’information généraliste (principalement Le Monde et Libération). Certains enquêtés ont par ailleurs été des fondateurs d’organes de presse pure players (Les Jours, Rue89, Slate.fr, Loopsider, Orient XXI…). En outre, d’autres enquêtés sont responsables de formation d’école de journalisme (Sciences Po, IPJ Dauphine, ESJ Lille) ou appartiennent à des associations professionnelles (Reporters sans frontières, Prenons la Une). Après un rappel sur le fonctionnement des algorithmes de recommandation et de leur portée, nous questionnons la manière dont ceux-ci mettent à mal le travail de production et de priorisation de l’information. Nous montrons également comment ces algorithmes révèlent et renforcent des pratiques et des tendances journalistiques socialement situées
Steady three-dimensional rotational flows: existence via Kato's approach to locally coercive problems
No full textStationary flows of an inviscid and incompressible fluid of constant density in the region , periodic in the second and third variables, are considered. The flux and the Bernoulli function are prescribed at each point of the boundary . The previous existence proof relying on the Nash-Moser iteration scheme is replaced by an adaptation of Kato's approach to locally coercive problems, allowing a more precise statement: the regularity required in Sobolev spaces is the one needed to ensure a basic local coercivity property, and there is a loss of control of only two derivatives in the obtained solutions. The underlying variational structure gives an additional property: the obtained solutions are local minimizers of an integral functional. The strategy of proof is first developed for a simpler nonlinear partial differential equation in two variables which satisfies a weaker form of ellipticity