Source author record

Gabor Csányi

Gabor Csányi appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

cond-mat.mtrl-sci cond-mat.supr-con Machine Learning physics.comp-ph

Catalog footprint

What is connected

2works

4topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2019arXiv

Machine Learning Inter-Atomic Potentials Generation Driven by Active Learning: A Case Study for Amorphous and Liquid Hafnium dioxide

We propose a novel active learning scheme for automatically sampling a minimum number of uncorrelated configurations for fitting the Gaussian Approximation Potential (GAP). Our active learning scheme consists of an unsupervised machine learning (ML) scheme coupled to Bayesian optimization technique that evaluates the GAP model. We apply this scheme to a Hafnium dioxide (HfO2) dataset generated from a melt-quench ab initio molecular dynamics (AIMD) protocol. Our results show that the active learning scheme, with no prior knowledge of the dataset is able to extract a configuration that reaches the required energy fit tolerance. Further, molecular dynamics (MD) simulations performed using this active learned GAP model on 6144-atom systems of amorphous and liquid state elucidate the structural properties of HfO2 with near ab initio precision and quench rates (i.e. 1.0 K/ps) not accessible via AIMD. The melt and amorphous x-ray structural factors generated from our simulation are in good agreement with experiment. Additionally, the calculated diffusion constants are in good agreement with previous ab initio studies.

preprint2005arXiv

Electronic structure of superconducting graphite intercalate compounds: The role of the interlayer state

Although not an intrinsic superconductor, it has been long--known that, when intercalated with certain dopants, graphite is capable of exhibiting superconductivity. Of the family of graphite--based materials which are known to superconduct, perhaps the most well--studied are the alkali metal--graphite intercalation compounds (GIC) and, of these, the most easily fabricated is the C${}_8$K system which exhibits a transition temperature $\bm{T_c\simeq 0.14} $K. By increasing the alkali metal concentration (through high pressure fabrication techniques), the transition temperature has been shown to increase to as much as $\bm 5 $K in C${}_2$Na. Lately, in an important recent development, Weller \emph{et al.} have shown that, at ambient conditions, the intercalated compounds \cyb and \cca exhibit superconductivity with transition temperatures $\bm{T_c\simeq 6.5} $K and $\bm{11.5} $K respectively, in excess of that presently reported for other graphite--based compounds. We explore the architecture of the states near the Fermi level and identify characteristics of the electronic band structure generic to GICs. As expected, we find that charge transfer from the intercalant atoms to the graphene sheets results in the occupation of the $\bmπ$--bands. Yet, remarkably, in all those -- and only those -- compounds that superconduct, we find that an interlayer state, which is well separated from the carbon sheets, also becomes occupied. We show that the energy of the interlayer band is controlled by a combination of its occupancy and the separation between the carbon layers.