arXiv subjects

Eli Meril

Publications and source records attributed to Eli Meril.

3 recordsLinked to original sources

arXiv2026

Mode Conversion of Gaussian Beams at Dielectric Interfaces

We investigate mode conversion of $\mathrm{TEM}_{00}$ Gaussian beams upon transmission through planar dielectric interfaces. We show that the angle-dependent Fresnel coefficients act as a spatial filter, inevitably generating higher-order spatial modes. Using a vector angular spectrum formulation and numerical simulations, we reveal that this polarization-dependent filtering induces a coupling from $\mathrm{TEM}_{00}$ into higher-order Laguerre-Gaussian modes, yielding a quadrupolar field pattern. We quantify the associated amplitude and phase deviations, showing that the mode fidelity decreases significantly as the beam waist approaches the diffraction limit.

physics.optics↗

arXiv2025

AlgoTune: Can Language Models Speed Up General-Purpose Numerical Programs?

Despite progress in language model (LM) capabilities, evaluations have thus far focused on models' performance on tasks that humans have previously solved, including in programming (Jimenez et al., 2024) and mathematics (Glazer et al., 2024). We therefore propose testing models' ability to design and implement algorithms in an open-ended benchmark: We task LMs with writing code that efficiently solves computationally challenging problems in computer science, physics, and mathematics. Our AlgoTune benchmark consists of 154 coding tasks collected from domain experts and a framework for validating and timing LM-synthesized solution code, which is compared to reference implementations from popular open-source packages. In addition, we develop a baseline LM agent, AlgoTuner, and evaluate its performance across a suite of frontier models. AlgoTuner uses a simple, budgeted loop that edits code, compiles and runs it, profiles performance, verifies correctness on tests, and selects the fastest valid version. AlgoTuner achieves an average 1.72x speedup against our reference solvers, which use libraries such as SciPy, sk-learn and CVXPY. However, we find that current models fail to discover algorithmic innovations, instead preferring surface-level optimizations. We hope that AlgoTune catalyzes the development of LM agents exhibiting creative problem solving beyond state-of-the-art human performance.

cs.SE↗

arXiv2025

Miniband Generation by Surface Acoustic Waves

We introduce a new class of tunable periodic structures, formed by launching two obliquely propagating surface acoustic waves on a piezoelectric substrate that supports a two-dimensional quantum material. The resulting acoustoelectric superlattice exhibits two salient features. First, its periodicity is widely tunable, spanning a length scale intermediate between moir\'e superlattices and optical lattices, enabling the formation of narrow, topologically nontrivial energy bands. Second, unlike moir\'e systems, where the superlattice amplitude is set by intrinsic interlayer tunneling and lattice relaxation, the amplitude of the acoustoelectric potential is externally tunable via the surface acoustic wave power. Using massive monolayer graphene as an example, we demonstrate that varying the frequencies and power of the surface acoustic waves enables in-situ control over the band structure of the 2D material, generating flat bands and nontrivial valley Chern numbers, featuring a highly localized Berry curvature.

cond-mat.mes-hall↗