4 papers · 1 filter
A single design choice determines whether machine learning models of materials make physically impossible predictions
Can Polat, Mustafa Kurban, Erchin Serpedin +1
Machine-learned models are replacing first-principles calculations across materials discovery, and physical symmetry is the central guarantee built into them. The debate over how m…
When do machine-learned exchange-correlation improvements inherit into density-functional tight binding?
Can Polat, Mustafa Kurban, Erchin Serpedin +1
Machine-learned exchange-correlation functionals correct band gaps at near-semilocal cost, while density-functional tight binding reaches the --atom regime; combining t…
How Far Can You Grow? Characterizing the Extrapolation Frontier of Graph Generative Models for Materials Science
Can Polat, Erchin Serpedin, Mustafa Kurban +1
Every generative model for crystalline materials harbors a critical structure size beyond which its outputs become unreliable; we call this the extrapolation frontier. Despite its…
C2NP: A Benchmark for Learning Scale-Dependent Geometric Invariances in 3D Materials Generation
Can Polat, Erchin Serpedin, Mustafa Kurban +1
Generative models for materials have achieved strong performance on periodic bulk crystals, yet their ability to generalize across scale transitions to finite nanostructures remain…