Laboratoire d'Analyse et d'Architecture des Systèmes

When Optimization Forgets the Minority: Bias Amplification in Policy-Gradient Reinforcement Learning

Bias amplification, the tendency of trained models to be more biased than the data that produced them, is well documented in supervised learning. We study it in policy-gradient reinforcement learning, where there is no static dataset: each episode belongs to a latent group that s...

Replicability of agent-based simulations: a case study with a SEIRS epidemiological model

Reproducibility is not only a cornerstone of the advancement of science, but also of scientific credibility. However, it remains a persistent challenge across many disciplines, and it is now referred to as the “reproducibility crisis.” This issue is particularly relevant in compu...

Scalable ground-state certification of quantum spin systems via structured noncommutative polynomial optimization

A fundamental challenge in quantum physics is determining the ground-state properties of many-body systems. Whereas standard variational approaches posit a wave-function ansatz and minimize over the possible states expressible by that ansatz, the problem can alternatively be form...

Grape yield estimation using on-the-fly MIMO millimeter-wave radar at operational field speed: A variety-dependent proof of concept

Accurate pre-harvest grape yield estimation is a critical challenge in precision viticulture, yet current optical approaches (RGB cameras, LiDAR) are fundamentally limited by foliage occlusion and sensitivity to ambient lighting conditions. This paper presents a ground-based grap...

Solving Stengle’s Example in Rational Arithmetic: Exact Values of the Moment-SOS Relaxations

We revisit Stengle’s classical univariate polynomial optimization example min 1 - x 2 s.t. ( 1 - x 2 ) 3 ≥ 0 whose constraint description is degenerate at the minimizers. We prove that the moment-SOS hierarchy of relaxation order r ≥ 3 has the exact value - 1 / ( r ( r - 2 ) ) ....

A predictive coding framework for safe and versatile control of supernumerary robotic limbs

Supernumerary robotic limbs (SLs) have the potential to extend human sensorimotor capabilities by increasing effective degrees of freedom. However, current systems are designed for a single task and cannot flexibly coordinate with natural limbs (NLs). Here we propose a hierarchic...

Extrinsic Limitations of Stealthy Hyperuniform Devices

ABSTRACT Hyperuniformity promises an unusual form of wave control: the suppression of elastic scattering over extended angular ranges without periodic order. Here, we present a comprehensive experimental and theoretical study of 2D stealthy hyperuniform metasurfaces operating at...

Convergence Rates for Polynomial Optimization on Set Products

Abstract. We consider polynomial optimization problems on Cartesian products of basic compact semialgebraic sets. The solution of such problems can be approximated as closely as desired by hierarchies of semidefinite programming relaxations, based on classical sums of squares cer...

GRB 260310A / SN 2026fgk: A Multi-Wavelength Study of a Nearby Underluminous Long GRB and SN with a Complex Afterglow

Abstract We present a comprehensive multi-wavelength study of GRB 260310A / SN 2026fgk, a nearby (z = 0.153), long-duration gamma-ray burst (GRB) with an exceptionally underluminous prompt γ-ray emission and a Comptonized spectrum. The burst occurred at the edge of a blue host ga...