École Polytechnique Fédérale de Lausanne

Gradient Descent Avoids Strict Saddles with a Simple Line-Search Method Too

It is known that gradient descent (GD) on a [Formula: see text] cost function generically avoids strict saddle points when using a small, constant step size. However, no such guarantee existed for GD with a line-search method. We provide one for a modified version of the standard...

Deciphering coccolith formation: advanced microscopy insights from the biomineralization of Gephyrocapsa huxleyi

Abstract Coccolithophores are unicellular marine phytoplankton that produce complex and intricately shaped mineralized scales called coccoliths. Coccoliths are produced in an intracellular vesicle where crystal nucleation occurs, from which several individual calcite units develo...

Associations between urban form and urban multifunctionality: a comparative study

Abstract Cities are complex and dynamic systems that integrate social, ecological, and infrastructure domains. This study investigates and quantifies how urban forms are associated with multiple urban functions across these domains, and how these associations combine into positiv...

4D reconstruction of alumina laser melt pools at 25 kHz via operando X-ray multi-projection imaging

Abstract Advancing additive manufacturing, e.g., laser powder-bed fusion (LPBF), requires resolving rapid processes such as melt-pool dynamics and keyhole evolution in 4D (3D + time). Operando X-ray tomography is a state-of-the-art approach for 4D characterization, but its tempor...

A User-Centered Approach to Customizing, Validating, and Evaluating a Large Language Model for Empowering Healthcare Providers with Up-to-Date Information on Adolescent and Child Health

Healthcare professionals (HCPs) in Kenya face critical challenges in accessing reliable, context-specific information on adolescent and child health, which hampers clinical decision-making and compromises patient outcomes. Large Language Models (LLMs) offer a promising solution b...

Thawed Gaussian Ehrenfest dynamics

Ehrenfest dynamics is a widely used mixed quantum-classical approach for nonadiabatic molecular dynamics, whereas thawed Gaussian wavepacket dynamics provides an efficient semiclassical description of adiabatic nuclear quantum dynamics. Here, we describe thawed Gaussian Ehrenfest...

Simulation-based deep learning for lightning localization using convolutional autoencoder and multilayer perceptron

Lightning-induced overvoltages are a major cause of faults and outages in overhead transmission systems, significantly affecting the reliability and operational continuity of power networks. Accurate and timely localization of lightning strikes is therefore essential for improvin...

Simultaneous T 1 and T 2 Mapping in the Brain With QuantoRAGE : An MP2RAGE Variant Using T 2 ‐Prepared Inversion

ABSTRACT Purpose To develop and validate a repeatable and reproducible approach, QuantoRAGE, for simultaneous whole‐brain T 1 and T 2 mapping using adiabatic magnetization preparation. Methods QuantoRAGE is a 3D FLASH‐based sequence using an adiabatic T 2 ‐prepared inversion foll...

Steric Stabilization of Lipid Nanoparticles by Sequence-Defined Polymer−Lipids Enhances Gene Expression

Abstract Polyethylene glycol (PEG) conjugated lipids are frequently used for the design of lipid nanoparticles (LNPs) in the field of gene therapy. However, conventional PEG−lipids contain a distribution of polymer chain lengths. To overcome this limitation, we have prepared a st...

MicroRNA 517a-3p suppresses α-synuclein fibril accumulation in human glioblastoma cells

Abstract Background Synucleinopathies are a group of neurodegenerative disorders characterized by the misfolding, aggregation, and accumulation of α-synuclein (α-syn) into oligomeric assemblies and amyloid inclusions. Alongside neurons, glial cells can internalize pathological α-...

Brief communication: A novel wake mixing phenomenon and key parameters for wake recovery of floating wind turbines subjected to surge motions

This letter clarifies the key parameters governing the wake recovery of a floating wind turbine undergoing surge motions. A dedicated wind tunnel campaign covering reduced frequencies ( St p ) well beyond the current literature is presented. This enabled a full characterisation o...

First-Class Refinement Types for Scala

Refinement types—types qualified with logical predicates—have proven effective for lightweight verification in languages like Liquid Haskell, F*, and Dafny. However, in these systems refinements are either written in a separate specification language or treated as second-class an...

Opening Pandora’s box: Paper mills in conference proceedings

OBJECTIVE: Paper mills are a growing threat to the integrity of science, yet their penetration in conference proceedings remains underexplored despite conferences being more important than journals in some scientific subfields. This study aims to identify papers in conference pro...

Variance reduction for forces and pressure in variational Monte Carlo

We present simple and practical strategies to reduce the variance of Monte Carlo estimators. Our focus is on variational Monte Carlo calculations of atomic forces and pressure in electronic systems, although we show that the underlying ideas apply more broadly to other observable...

Rectifying market externalities via AI policymaking

Abstract Traditional competitive markets do not account for negative externalities; indirect costs that some participants impose on others, such as the cost of over-appropriating a common-pool resource (which diminishes future stock, and thus harvest, for everyone). Quantifying a...

Hierarchical quasiparticle dynamics in antiferromagnets revealed by time- and momentum-resolved X-ray scattering

Abstract Energy flows among coupled subsystems are essential for ultrafast dynamics and high-speed technologies. In magnetic materials, spin fluctuations —magnons— mediate these flows in ultrafast magnetism. Yet momentum-resolved access to low-energy magnons governing the microsc...

An Arctic climatology of particle number size distributions and cloud condensation nuclei during warm and moist intrusions

Abstract Recent studies indicate that warm and moist intrusions are significant sources of aerosols in the Arctic. Yet their contribution to different aerosol particle size modes, cloud condensation nuclei (CCN), and droplet number concentrations ( N d ) remains largely unconstra...

MEGATRON: how the first stars can create an iron metallicity plateau in the smallest dwarf galaxies

We study the stellar mass-iron metallicity relation of dwarf galaxies in the new high-resolution MEGATRON cosmological radiation-hydrodynamics simulations. These simulations model galaxy formation up to z ≈ 8 in a region that will collapse into a Milky-Way-like galaxy at z = 0 ,...

Multiscale transitional flow in anisotropic nanoparticle suspensions revealed by time-resolved X-ray scatter microscopy

Abstract Complex fluids transition from laminar to transitional flow above a critical control parameter, in a similar manner to their Newtonian counterparts. In a continuum mechanics sense, fluid elements follow the ensuing complex trajectories, which gives rise to secondary flow...

Emergent cavity-QED dynamics along the edge of a photonic lattice

Abstract We investigate qubits coupled to the boundary of a two-dimensional photonic lattice that supports dispersionless edge modes—unlike conventional edge modes that sustain propagating photons. As a case study, we consider a honeycomb lattice ( photonic graphene ) of coupled...