arXiv subjects

Roland Bouffanais

Publications and source records attributed to Roland Bouffanais.

At least 19 recordsLinked to original sources

arXiv2026

IntLawNER: A Named Entity Recognition Dataset and Benchmark in International Law

International law provides the normative framework through which states coordinate action, regulate armed conflict, and protect human rights, yet its texts remain without token-level named entity recognition (NER) resources. We introduce IntLawNER, a NER dataset and benchmark for codified sources of international law, covering 2,987 gold-annotated sentences and 8,094 entity spans from International Court of Justice (ICJ) decisions, UN Security Council resolutions, and European Court of Human Rights (ECtHR) judgments, annotated with seven institution-specific entity types. We construct IntLawNER with a cost-effective hybrid algorithmic-agentic pipeline that reduces 468k source sentences to a compact annotation set through candidate retrieval, LLM-based vetting, and human review, with 89.6% of gold spans accepted unchanged from the silver layer. However, the silver-to-gold analysis reveals that human-machine aggregate agreement metrics can be misleading in domain-specific NER: Cohen's kappa=0.964 on boundary-matched spans masks a macro-F1 of 0.753 when missing entities, boundary errors, and label corrections are included. The benchmark shows that zero-shot span-based GLiNER collapses on entity types dependent on institutional function rather than surface form (0.243 micro-F1), while fine-tuned transformers struggle on rare labels. Carefully selected few-shot examples that demonstrate label contrasts improve every LLM over zero-shot prompting, with Claude Opus 4.6 reaching the best score of 0.873 micro-F1. We release IntLawNER as a benchmark and reusable resource for extracting references in international legal texts.

cs.AI↗

arXiv2026

Aligning the Unseen in Attributed Graphs: Interplay between Graph Geometry and Node Attributes Manifold

The standard approach to representation learning on attributed graphs -- i.e., simultaneously reconstructing node attributes and graph structure -- is geometrically flawed, as it merges two potentially incompatible metric spaces. This forces a destructive alignment that erodes information about the graph's underlying generative process. To recover this lost signal, we introduce a custom variational autoencoder that separates manifold learning from structural alignment. By quantifying the metric distortion needed to map the attribute manifold onto the graph's Heat Kernel, we transform geometric conflict into an interpretable structural descriptor. Experiments show our method uncovers connectivity patterns and anomalies undetectable by conventional approaches, proving both their theoretical inadequacy and practical limitations.

cs.AI↗

arXiv2025

Revealing Higher-Order Interactions in Complex Networks: A U.S. Diplomacy Case Study

Although diplomatic communication has long been examined in the social sciences, its network structure remains underexplored. Using the U.S. diplomatic cables released by WikiLeaks in 2010 as a case study, we adopt a network-science perspective. We represent diplomatic interactions as a hypergraph and develop a general, random-walk-based pipeline to evaluate this representation against traditional pairwise graphs. We further evaluate the pipeline on legislative co-sponsorship and organizational email data, finding improvements and empirical evidence that clarifies when hypergraph modeling is preferable to pairwise graphs. Overall, hypergraphs paired with appropriately specified random-walk dynamics more faithfully capture higher-order, group-based interactions, yielding a richer structural account of diplomacy and superior performance on interaction-prediction tasks that enables inferring new diplomatic relationships from existing patterns.

cs.SI↗

arXiv2023

A Sequential Deep Learning Algorithm for Sampled Mixed-integer Optimisation Problems

Mixed-integer optimisation problems can be computationally challenging. Here, we introduce and analyse two efficient algorithms with a specific sequential design that are aimed at dealing with sampled problems within this class. At each iteration step of both algorithms, we first test the feasibility of a given test solution for each and every constraint associated with the sampled optimisation at hand, while also identifying those constraints that are violated. Subsequently, an optimisation problem is constructed with a constraint set consisting of the current basis -- namely, the smallest set of constraints that fully specifies the current test solution -- as well as constraints related to a limited number of the identified violating samples. We show that both algorithms exhibit finite-time convergence towards the optimal solution. Algorithm 2 features a neural network classifier that notably improves the computational performance compared to Algorithm 1. We quantitatively establish these algorithms' efficacy through three numerical tests: robust optimal power flow, robust unit commitment, and robust random mixed-integer linear program.

math.OC↗

arXiv2023

Multi-Target Pursuit by a Decentralized Heterogeneous UAV Swarm using Deep Multi-Agent Reinforcement Learning

Multi-agent pursuit-evasion tasks involving intelligent targets are notoriously challenging coordination problems. In this paper, we investigate new ways to learn such coordinated behaviors of unmanned aerial vehicles (UAVs) aimed at keeping track of multiple evasive targets. Within a Multi-Agent Reinforcement Learning (MARL) framework, we specifically propose a variant of the Multi-Agent Deep Deterministic Policy Gradient (MADDPG) method. Our approach addresses multi-target pursuit-evasion scenarios within non-stationary and unknown environments with random obstacles. In addition, given the critical role played by collective exploration in terms of detecting possible targets, we implement heterogeneous roles for the pursuers for enhanced exploratory actions balanced by exploitation (i.e. tracking) of previously identified targets. Our proposed role-based MADDPG algorithm is not only able to track multiple targets, but also is able to explore for possible targets by means of the proposed Voronoi-based rewarding policy. We implemented, tested and validated our approach in a simulation environment prior to deploying a real-world multi-robot system comprising of Crazyflie drones. Our results demonstrate that a multi-agent pursuit team has the ability to learn highly efficient coordinated control policies in terms of target tracking and exploration even when confronted with multiple fast evasive targets in complex environments.

cs.RO↗

arXiv2023

Effect of Swarm Density on Collective Tracking Performance

How does the size of a swarm affect its collective action? Despite being arguably a key parameter, no systematic and satisfactory guiding principles exist to select the number of units required for a given task and environment. Even when limited by practical considerations, system designers should endeavor to identify what a reasonable swarm size should be. Here, we show that this fundamental question is closely linked to that of selecting an appropriate swarm density. Our analysis of the influence of density on the collective performance of a target tracking task reveals different `phases' corresponding to markedly distinct group dynamics. We identify a `transition' phase, in which a complex emergent collective response arises. Interestingly, the collective dynamics within this transition phase exhibit a clear trade-off between exploratory actions and exploitative ones. We show that at any density, the exploration-exploitation balance can be adjusted to maximize the system's performance through various means, such as by changing the level of connectivity between agents. While the density is the primary factor to be considered, it should not be the sole one to be accounted for when sizing the system. Due to the inherent finite-size effects present in physical systems, we establish that the number of constituents primarily affects system-level properties such as exploitation in the transition phase. These results illustrate that instead of learning and optimizing a swarm's behavior for a specific set of task parameters, further work should instead concentrate on learning to be adaptive, thereby endowing the swarm with the highly desirable feature of being able to operate effectively over a wide range of circumstances.

cs.MA↗

arXiv2022

Controllability of a Class of Swarm Signaling Networks

In this paper, we propose closed-form analytical expressions to determine the minimum number of driver nodes that is needed to control a specific class of networks. We consider swarm signaling networks with regular out-degree distribution where a fraction $p$ of the links is unavailable. We further apply our method to networks with bi-modal out-degree distributions. Our approximations are validated through intensive simulations. Results show that our approximations have high accuracy when compared with simulation results for both types of out-degree distribution.

eess.SY↗

arXiv2022

Adapting the Exploration-Exploitation Balance in Heterogeneous Swarms: Tracking Evasive Targets

There has been growing interest in the use of multi-robot systems in various tasks and scenarios. The main attractiveness of such systems is their flexibility, robustness, and scalability. An often overlooked yet promising feature is system modularity, which offers the possibility to harness agent specialization, while also enabling system-level upgrades. However, altering the agents' capacities can change the exploration-exploitation balance required to maximize the system's performance. Here, we study the effect of a swarm's heterogeneity on its exploration-exploitation balance while tracking multiple fast-moving evasive targets under the Cooperative Multi-Robot Observation of Multiple Moving Targets framework. To this end, we use a decentralized search and tracking strategy with adjustable levels of exploration and exploitation. By indirectly tuning the balance, we first confirm the presence of an optimal balance between these two key competing actions. Next, by substituting slower moving agents with faster ones, we show that the system exhibits a performance improvement without any modifications to the original strategy. In addition, owing to the additional amount of exploitation carried out by the faster agents, we demonstrate that a heterogeneous system's performance can be further improved by reducing an agent's level of connectivity, to favor the conduct of exploratory actions. Furthermore, in studying the influence of the density of swarming agents, we show that the addition of faster agents can counterbalance a reduction in the overall number of agents while maintaining the level of tracking performance. Finally, we explore the challenges of using differentiated strategies to take advantage of the heterogeneous nature of the swarm.

cs.RO↗

arXiv2021

Tracking Multiple Fast Targets With Swarms: Interplay Between Social Interaction and Agent Memory

The task of searching for and tracking of multiple targets is a challenging one. However, most works in this area do not consider evasive targets that move faster than the agents comprising the multi-robot system. This is due to the assumption that the movement patterns of such targets, combined with their excessive speed, would make the task nearly impossible to accomplish. In this work, we show that this is not the case and we propose a decentralized search and tracking strategy in which the level of exploration and exploitation carried out by the swarm is adjustable. By tuning a swarm's exploration and exploitation dynamics, we demonstrate that there exists an optimal balance between the level of exploration and exploitation performed. This optimum maximizes its tracking performance and changes depending on the number of targets and the targets' movement profiles. We also show that the use of agent-based memory is critical in enabling the tracking of an evasive target. The obtained simulation results are validated through experimental tests with a decentralized swarm of six robots tracking a virtual fast-moving target.

cs.MA↗

arXiv2021

Tuning the Clustering Coefficient of Generalized Circulant Networks

Apart from the role the clustering coefficient plays in the definition of the small-world phenomena, it also has great relevance for practical problems involving networked dynamical systems. To study the impact of the clustering coefficient on dynamical processes taking place on networks, some authors have focused on the construction of graphs with tunable clustering coefficients. These constructions are usually realized through a stochastic process, either by growing a network through the preferential attachment procedure, or by applying a random rewiring process. In contrast, we consider here several families of static graphs whose clustering coefficients can be determined explicitly. The basis for these families is formed by the $k$-regular graphs on $N$ nodes, that belong to the family of so-called circulant graphs denoted by $C_{N,k}$. We show that the expression for the clustering coefficient of $C_{N,k}$ reported in literature, only holds for sufficiently large $N$. Next, we consider three generalizations of the circulant graphs, either by adding some pendant links to $C_{N,k}$, or by connecting, in two different ways, an additional node to some nodes of $C_{N,k}$. For all three generalizations, we derive explicit expressions for the clustering coefficient. Finally, we construct a family of pairs of generalized circulant graphs, with the same number of nodes and links, but with different clustering coefficients.

physics.soc-ph↗

arXiv2021

Perturbation symmetries in shear-thinning viscoelastic pipe flows and the Petrov-Galerkin implementation

The perturbations of the laminar shear-thinning viscoelastic pipe flow under Finitely Extensible Nonlinear Elastic model with Peterlin approximation (FENE-P) are shown to exhibit leading-order power-law behaviours, and the expected odd-even parities with respect to the radial coordinate that depend on the azimuthal wave\-number, $n$. The analysis helps regularizing the governing system of equations at the centreline, and allows for a complete stability analysis of three-dimensional perturbations for a general integer value of $n$, which has hitherto remained a challenge for FENE-P models. It is shown here that the symmetry and analytic behaviours of the velocity and pressure fields of the Newtonian counterpart are both preserved in this flow, and the reason is elucidated. For $|n|=1$, the perturbations to the correlations between the axial component and the radial or azimuthal components of the end-to-end polymer vector exhibit behaviour similar to that of the velocity perturbations close to the centreline, and are traced to the uniformity of axial traction with respect to the azimuthal direction. For all values of $n$, the fluctuation to the end-to-end length of the polymer chain vanishes at the centreline. The ansatzes for the perturbations to the components of conformation tensor arrived here, using heuristics, are later proved in two separate ways: Frobenius method, and a method that utilizes observations from Fourier analysis. We also reveal the existence of natural modes of polymers with trivial velocity perturbations in the limit of vanishing polymer viscosity. The complex frequency spectrum of these modes is continuous with radial stratification. Finally, the ansatzes are implemented using a Petrov--Galerkin spectral scheme.

physics.flu-dyn↗

arXiv2021

External light control of three-dimensional ultrashort far-infrared pulses in an inhomogeneous array of carbon nanotubes

We present a study of the propagation of three-dimensional (3D) bipolar electromagnetic ultrashort pulses in an inhomogeneous array of semiconductor carbon nanotubes (CNTs) in the presence of a control high-frequency (HF) electric field. The inhomogeneity is present in the form of a layer with an increased concentration of conduction electrons, which acts as a barrier for the propagation of ultrashort electromagnetic pulses through the CNT array. The dynamics of the pulse is described by a nonlinear equation for the vector potential of the electromagnetic field (it takes the form of a 3D generalization of the sine-Gordon equation), derived from the Maxwell's equations and averaged over the period of the HF control field. By means of systematic simulations, we demonstrate that, depending on the amplitude and frequency of the HF control, the ultrashort pulse approaching the barrier layer either passes it or bounces back. The layer's transmissivity for the incident pulse is significantly affected by the amplitude and frequency of the HF control field, with the reflection coefficient nearly vanishing in intervals that make up a discrete set of transparency windows, which resembles the effect of the electromagnetically-induced transparency. Having passed the barrier, the ultrashort pulse continues to propagate, keeping its spatiotemporal integrity. The results may be used for the design of soliton valves, with the transmissivity of the soliton stream accurately controlled by the HF field.

physics.optics↗

arXiv2020

Multi-Agent Reinforcement Learning for Dynamic Ocean Monitoring by a Swarm of Buoys

Autonomous marine environmental monitoring problem traditionally encompasses an area coverage problem which can only be effectively carried out by a multi-robot system. In this paper, we focus on robotic swarms that are typically operated and controlled by means of simple swarming behaviors obtained from a subtle, yet ad hoc combination of bio-inspired strategies. We propose a novel and structured approach for area coverage using multi-agent reinforcement learning (MARL) which effectively deals with the non-stationarity of environmental features. Specifically, we propose two dynamic area coverage approaches: (1) swarm-based MARL, and (2) coverage-range-based MARL. The former is trained using the multi-agent deep deterministic policy gradient (MADDPG) approach whereas, a modified version of MADDPG is introduced for the latter with a reward function that intrinsically leads to a collective behavior. Both methods are tested and validated with different geometric shaped regions with equal surface area (square vs. rectangle) yielding acceptable area coverage, and benefiting from the structured learning in non-stationary environments. Both approaches are advantageous compared to a naïve swarming method. However, coverage-range-based MARL outperforms the swarm-based MARL with stronger convergence features in learning criteria and higher spreading of agents for area coverage.

cs.RO↗

arXiv2020

Heterogeneous Swarms for Maritime Dynamic Target Search and Tracking

Current strategies employed for maritime target search and tracking are primarily based on the use of agents following a predetermined path to perform a systematic sweep of a search area. Recently, dynamic Particle Swarm Optimization (PSO) algorithms have been used together with swarming multi-robot systems (MRS), giving search and tracking solutions the added properties of robustness, scalability, and flexibility. Swarming MRS also give the end-user the opportunity to incrementally upgrade the robotic system, inevitably leading to the use of heterogeneous swarming MRS. However, such systems have not been well studied and incorporating upgraded agents into a swarm may result in degraded mission performances. In this paper, we propose a PSO-based strategy using a topological k-nearest neighbor graph with tunable exploration and exploitation dynamics with an adaptive repulsion parameter. This strategy is implemented within a simulated swarm of 50 agents with varying proportions of fast agents tracking a target represented by a fictitious binary function. Through these simulations, we are able to demonstrate an increase in the swarm's collective response level and target tracking performance by substituting in a proportion of fast buoys.

cs.RO↗

arXiv2020

Spatial super-spreaders and super-susceptibles in human movement networks

As lockdowns and stay-at-home orders start to be lifted across the globe, governments are struggling to establish effective and practical guidelines to reopen their economies. In dense urban environments with people returning to work and public transportation resuming full capacity, enforcing strict social distancing measures will be extremely challenging, if not practically impossible. Governments are thus paying close attention to particular locations that may become the next cluster of disease spreading. Indeed, certain places, like some people, can be "super-spreaders." Is a bustling train station in a central business district more or less susceptible and vulnerable as compared to teeming bus interchanges in the suburbs? Here, we propose a quantitative and systematic framework to identify spatial super-spreaders and the novel concept of super-susceptibles, i.e. respectively, places most likely to contribute to disease spread or to people contracting it. Our proposed data-analytic framework is based on the daily-aggregated ridership data of public transport in Singapore. By constructing the directed and weighted human movement networks and integrating human flow intensity with two neighborhood diversity metrics, we are able to pinpoint super-spreader and super-susceptible locations. Our results reveal that most super-spreaders are also super-susceptibles and that counterintuitively, busy peripheral bus interchanges are riskier places than crowded central train stations. Our analysis is based on data from Singapore, but can be readily adapted and extended for any other major urban center. It therefore serves as a useful framework for devising targeted and cost-effective preventive measures for urban planning and epidemiological preparedness.

physics.soc-ph↗

arXiv2020

Viscoelastic laminar drag bounds in pipe flow

The velocity and friction properties of laminar pipe flow of a viscoelastic solution are bounded by the corresponding values for two Newtonian fluids, namely, the solvent and a fluid with a viscosity identical to the total viscosity of the solution. The lower friction factor for the flow of the solution when compared to the latter is tracked to an increased strain rate needed to enhance viscous dissipation. Lastly, we show analytically that the effective viscosity varies similarly to the radial diagonal component of the conformation tensor as observed numerically in turbulent flows, and give a lucid interpretation of shear-thinning through a sequence of underlying constitutive physical phenomena.

physics.flu-dyn↗

arXiv2019

A Physics-Based Attack Detection Technique in Cyber-Physical Systems: A Model Predictive Control Co-Design Approach

In this paper a novel approach to co-design controller and attack detector for nonlinear cyber-physical systems affected by false data injection (FDI) attack is proposed. We augment the model predictive controller with an additional constraint requiring the future---in some steps ahead---trajectory of the system to remain in some time-invariant neighborhood of a properly designed reference trajectory. At any sampling time, we compare the real-time trajectory of the system with the designed reference trajectory, and construct a residual. The residual is then used in a nonparametric cumulative sum (CUSUM) anomaly detector to uncover FDI attacks on input and measurement channels. The effectiveness of the proposed approach is tested with a nonlinear model regarding level control of coupled tanks.

eess.SY↗

arXiv2019

From Senseless Swarms to Smart Mobs: Tuning Networks for Prosocial Behaviour

Social media have been seen to accelerate the spread of negative content such as disinformation and hate speech, often unleashing reckless herd mentality within networks, further aggravated by malicious entities using bots for amplification. So far, the response to this emerging global crisis has centred around social media platform companies making reactive moves that appear to have greater symbolic value than practical utility. These include taking down patently objectionable content or manually deactivating the accounts of bad actors, while leaving vast troves of negative content to circulate and perpetuate within social networks. Governments worldwide have thus sought to intervene using regulatory tools, with countries such as France, Germany and Singapore introducing laws to compel technology companies to take down or correct erroneous and harmful content. However, the relentless pace of technological progress enfeebles regulatory measures that seem fated for obsolescence.

cs.SI↗