Results 161 to 170 of about 139,835 (302)
On the Almost Sure Convergence of Stochastic Gradient Descent in Non-Convex Problems [PDF]
This paper analyzes the trajectories of stochastic gradient descent (SGD) to help understand the algorithm’s convergence properties in non-convex problems.
Mertikopoulos, Panayotis +3 more
core
How Long does it Take to Train an Elephant Random Walk. [PDF]
Fang Z.
europepmc +1 more source
A Practical Noise2Noise Denoising Pipeline for High‐Throughput Raman Spectroscopy
A lightweight and reproducible denoising pipeline for high‐throughput Raman spectroscopy is introduced, based on a 1D convolutional autoencoder trained with a Noise2Noise strategy. Using only repeated short‐exposure acquisitions, the method suppresses stochastic noise without reference spectra, enabling reliable spectral reconstruction while preserving
David Martin‐Calle +5 more
wiley +1 more source
High-Contrast Random Composites: Homogenisation Framework and Spectral Convergence. [PDF]
Cherdantsev M +2 more
europepmc +1 more source
Electrified Damage in Motion Systems
The electrified damage in motion systems is a fundamental framework presenting the degradation pathway arising from the coupling of electrical energy transport with mechanical contact and interfacial chemistry. The framework positions electrified damage as a distinct degradation regime with unique characteristic surface morphologies and failures of ...
M. Humaun Kabir +2 more
wiley +1 more source
Almost Sure Convergence to Zero in Stochastic Growth Models [PDF]
This paper shows that in stochastic one-sector growth models, if the production function does not satisfy the Inada condition at zero, any feasible path converges to zero with probability one provided that the shocks are sufficiently volatile.
Takashi Kamihigashi
core
Varifold solutions to volume-preserving mean curvature flow: existence and weak-strong uniqueness. [PDF]
Poiatti A.
europepmc +1 more source
Design and Analysis of Compression–Torsion Coupling Metamaterials Using the Golden Section Method
A novel compression–torsion metamaterial is engineered using inclined rods and symmetry breaking. To optimize its torsional performance, the golden section method is employed. The mechanical response of the metamaterial is validated through both numerical analysis and experimental validation.
Amirhossein Hassani, Sara Bagherifard
wiley +1 more source
Reinforcement learning-based optimal control for stochastic opinion dynamics. [PDF]
Chen Y, Gao H, Mazalov VV, Liu Y.
europepmc +1 more source
Experiments and thermophysical simulations were conducted to investigate the electron beam powder bed fusion electron beam (PBF‐EB/M) process for the γ′‐strengthened nickel‐based superalloy Inconel 738LC. The results demonstrate the impact of process‐induced microstructural variations on high‐temperature mechanical behavior, providing a basis for ...
Jan Niklas Petenati +11 more
wiley +1 more source

