Results 181 to 190 of about 38,756 (312)
A Comprehensive Assessment and Benchmark Study of Large Atomistic Foundation Models for Phonons
We benchmark six large atomistic foundation models on 2429 crystalline materials for phonon transport properties. The rapid development of universal machine learning potentials (uMLPs) has enabled efficient, accurate predictions of diverse material properties across broad chemical spaces.
Md Zaibul Anam +5 more
wiley +1 more source
Data-driven regression analysis of amylose using Sombor molecular descriptors. [PDF]
Mufti ZS +4 more
europepmc +1 more source
Chat computational fluid dynamics (CFD) introduces an large language model (LLM)‐driven agent that automates OpenFOAM simulations end‐to‐end, attaining 82.1% execution success and 68.12% physical fidelity across 315 benchmarks—far surpassing prior systems.
E Fan +8 more
wiley +1 more source
ShockHash: Near Optimal-Space Minimal Perfect Hashing Beyond Brute-Force. [PDF]
Lehmann HP, Sanders P, Walzer S.
europepmc +1 more source
A low‐cost, self‐driving laboratory is developed to democratize autonomous materials discovery. Using this "frugal twin" hardware architecture with Bayesian optimization, the platform rapidly converges to target lower critical solution temperature (LCST) values while self‐correcting from off‐target experiments, demonstrating an accessible route to data‐
Guoyue Xu, Renzheng Zhang, Tengfei Luo
wiley +1 more source
Leveraging Different Distance Functions to Predict Antiviral Peptides with Geometric Deep Learning from ESMFold-Predicted Tertiary Structures. [PDF]
Cordoves-Delgado G +4 more
europepmc +1 more source
It is a fact that slippage causes tracking errors in both longitudinal and lateral directions which results to have less travel distance in tracking a reference trajectory. Less travel distance means having energy loss of the battery and carrying loads less than planned.
Gokhan Bayar +2 more
wiley +1 more source
This study presents a multitask strategy for plastic cleanup with autonomous surface vehicles, combining exploration and cleaning phases. A two‐headed Deep Q‐Network shared by all agents is traineded via multiobjective reinforcement learning, producing a Pareto front of trade‐offs.
Dame Seck +4 more
wiley +1 more source
PG-SCUnK: measuring pangenome graph representativeness using single-copy and universal K-mers. [PDF]
Cumer T, Milia S, Leonard AS, Pausch H.
europepmc +1 more source

