2026
|
Lazarev, Mikhail; Ustyuzhanin, Andrey Symbolic regression for defect interactions in 2D materials MATERIALS & DESIGN, 264 , 2026, DOI: 10.1016/j.matdes.2026.115706. Abstract | BibTeX | Endnote @article{WOS:001705842800001,
title = {Symbolic regression for defect interactions in 2D materials},
author = {Mikhail Lazarev and Andrey Ustyuzhanin},
doi = {10.1016/j.matdes.2026.115706},
times_cited = {0},
issn = {0264-1275},
year = {2026},
date = {2026-04-01},
journal = {MATERIALS & DESIGN},
volume = {264},
publisher = {ELSEVIER SCI LTD},
address = {125 London Wall, London, ENGLAND},
abstract = {Machine learning models have become firmly established across all
scientific fields. Extracting features from data and making inferences
based on them with neural network models often yields high accuracy;
however, this approach has several drawbacks. Symbolic regression is a
powerful technique for discovering analytical equations that describe
data, providing interpretable and generalizable models capable of
predicting unseen data. Symbolic regression methods have gained new
momentum with the advancement of neural network technologies and offer
several advantages, the main one being the interpretability of results.
In this work, we examined the application of the deep symbolic
regression algorithm SEGVAE to determine the properties of
two-dimensional materials with defects. Comparing the results with
state-of-the-art graph neural network-based methods shows comparable or,
in some cases, even identical outcomes. We also discuss the
applicability of this class of methods in natural sciences.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Machine learning models have become firmly established across all
scientific fields. Extracting features from data and making inferences
based on them with neural network models often yields high accuracy;
however, this approach has several drawbacks. Symbolic regression is a
powerful technique for discovering analytical equations that describe
data, providing interpretable and generalizable models capable of
predicting unseen data. Symbolic regression methods have gained new
momentum with the advancement of neural network technologies and offer
several advantages, the main one being the interpretability of results.
In this work, we examined the application of the deep symbolic
regression algorithm SEGVAE to determine the properties of
two-dimensional materials with defects. Comparing the results with
state-of-the-art graph neural network-based methods shows comparable or,
in some cases, even identical outcomes. We also discuss the
applicability of this class of methods in natural sciences. - FNClarivate Analytics Web of Science
- VR1.0
- PTJ
- AFMikhail Lazarev
Andrey Ustyuzhanin
- TISymbolic regression for defect interactions in 2D materials
- SOMATERIALS & DESIGN
- DTArticle
- ABMachine learning models have become firmly established across all
scientific fields. Extracting features from data and making inferences
based on them with neural network models often yields high accuracy;
however, this approach has several drawbacks. Symbolic regression is a
powerful technique for discovering analytical equations that describe
data, providing interpretable and generalizable models capable of
predicting unseen data. Symbolic regression methods have gained new
momentum with the advancement of neural network technologies and offer
several advantages, the main one being the interpretability of results.
In this work, we examined the application of the deep symbolic
regression algorithm SEGVAE to determine the properties of
two-dimensional materials with defects. Comparing the results with
state-of-the-art graph neural network-based methods shows comparable or,
in some cases, even identical outcomes. We also discuss the
applicability of this class of methods in natural sciences. - Z90
- PUELSEVIER SCI LTD
- PA125 London Wall, London, ENGLAND
- SN0264-1275
- VL264
- DI10.1016/j.matdes.2026.115706
- UTWOS:001705842800001
- ER
- EF
|
Maevskiy, Artem; Kapitan, Vitalii; Ustyuzhanin, Andrey Artificial Intelligence for Multiscale Modeling in Solid-State Physics
and Chemistry: A Comprehensive Review ADVANCED INTELLIGENT SYSTEMS, 2026, DOI: 10.1002/aisy.202501219. Abstract | BibTeX | Endnote @article{WOS:001711023100001,
title = {Artificial Intelligence for Multiscale Modeling in Solid-State Physics
and Chemistry: A Comprehensive Review},
author = {Artem Maevskiy and Vitalii Kapitan and Andrey Ustyuzhanin},
doi = {10.1002/aisy.202501219},
times_cited = {0},
year = {2026},
date = {2026-03-01},
journal = {ADVANCED INTELLIGENT SYSTEMS},
publisher = {WILEY-V C H VERLAG GMBH},
address = {POSTFACH 101161, 69451 WEINHEIM, GERMANY},
abstract = {Recent progress in artificial intelligence (AI) has transformed
methodologies across many areas of science. In materials research, AI
has enabled more efficient multiscale modeling by linking atomic,
mesoscale, and continuum scales with improved accuracy and reduced
computational cost. This review examines AI-based approaches in this
context and discusses their relationship to conventional analytical and
computational multiscale methods. Developments such as machine learning
force fields, graph neural networks, and AI-accelerated electronic
structure prediction are assessed with respect to their capabilities and
limitations. To illustrate the current state of the art in this field,
available software, computational tools, and benchmarks are discussed.
Applications in areas such as phase transitions, defect dynamics, and
bulk property prediction are shown, with an emphasis on how AI enhances
predictive capabilities. While highlighting the above-mentioned recent
advances, existing challenges and promising directions are also
discussed. This review is intended for two audiences: For AI
researchers, it demonstrates how physical and chemical constraints
influence models' development to ensure physical consistency, and for
physicists, chemists, and materials scientists, it illustrates how AI
can improve multiscale methods to solve previously inaccessible problems},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Recent progress in artificial intelligence (AI) has transformed
methodologies across many areas of science. In materials research, AI
has enabled more efficient multiscale modeling by linking atomic,
mesoscale, and continuum scales with improved accuracy and reduced
computational cost. This review examines AI-based approaches in this
context and discusses their relationship to conventional analytical and
computational multiscale methods. Developments such as machine learning
force fields, graph neural networks, and AI-accelerated electronic
structure prediction are assessed with respect to their capabilities and
limitations. To illustrate the current state of the art in this field,
available software, computational tools, and benchmarks are discussed.
Applications in areas such as phase transitions, defect dynamics, and
bulk property prediction are shown, with an emphasis on how AI enhances
predictive capabilities. While highlighting the above-mentioned recent
advances, existing challenges and promising directions are also
discussed. This review is intended for two audiences: For AI
researchers, it demonstrates how physical and chemical constraints
influence models' development to ensure physical consistency, and for
physicists, chemists, and materials scientists, it illustrates how AI
can improve multiscale methods to solve previously inaccessible problems - FNClarivate Analytics Web of Science
- VR1.0
- PTJ
- AFArtem Maevskiy
Vitalii Kapitan
Andrey Ustyuzhanin
- TIArtificial Intelligence for Multiscale Modeling in Solid-State Physics
and Chemistry: A Comprehensive Review - SOADVANCED INTELLIGENT SYSTEMS
- DTArticle
- ABRecent progress in artificial intelligence (AI) has transformed
methodologies across many areas of science. In materials research, AI
has enabled more efficient multiscale modeling by linking atomic,
mesoscale, and continuum scales with improved accuracy and reduced
computational cost. This review examines AI-based approaches in this
context and discusses their relationship to conventional analytical and
computational multiscale methods. Developments such as machine learning
force fields, graph neural networks, and AI-accelerated electronic
structure prediction are assessed with respect to their capabilities and
limitations. To illustrate the current state of the art in this field,
available software, computational tools, and benchmarks are discussed.
Applications in areas such as phase transitions, defect dynamics, and
bulk property prediction are shown, with an emphasis on how AI enhances
predictive capabilities. While highlighting the above-mentioned recent
advances, existing challenges and promising directions are also
discussed. This review is intended for two audiences: For AI
researchers, it demonstrates how physical and chemical constraints
influence models' development to ensure physical consistency, and for
physicists, chemists, and materials scientists, it illustrates how AI
can improve multiscale methods to solve previously inaccessible problems - Z90
- PUWILEY-V C H VERLAG GMBH
- PAPOSTFACH 101161, 69451 WEINHEIM, GERMANY
- DI10.1002/aisy.202501219
- UTWOS:001711023100001
- ER
- EF
|
2025
|
Lukianov, M Y; Maevskiy, A; Kazeev, N; Mylnikov, D; Svintsov, D A; Novoselov, K S; Ustyuzhanin, A; Bandurin, D A Inverse design of broadband antennas for terahertz devices based on
two-dimensional materials PHYSICAL REVIEW APPLIED, 24 (5), 2025, DOI: 10.1103/gr2z-3qjp. Abstract | BibTeX | Endnote @article{WOS:001633477300003,
title = {Inverse design of broadband antennas for terahertz devices based on
two-dimensional materials},
author = {M Y Lukianov and A Maevskiy and N Kazeev and D Mylnikov and D A Svintsov and K S Novoselov and A Ustyuzhanin and D A Bandurin},
doi = {10.1103/gr2z-3qjp},
times_cited = {0},
issn = {2331-7019},
year = {2025},
date = {2025-11-01},
journal = {PHYSICAL REVIEW APPLIED},
volume = {24},
number = {5},
publisher = {AMER PHYSICAL SOC},
address = {ONE PHYSICS ELLIPSE, COLLEGE PK, MD 20740-3844 USA},
abstract = {Terahertz technology, a cornerstone of next-generation high-speed
communication and sensing, has long been hindered by impedance mismatch
challenges that limit device performance and applicability. These
challenges become particularly pronounced when ultrasensitive
two-dimensional (2D) materials are employed as the detecting element in
the terahertz range, further complicating their integration in realworld
applications. Furthermore, conventional antenna designs often fail to
provide adequate matching across the broad terahertz spectrum. In this
work, we tackle these challenges using a procedural generation algorithm
to design terahertz broadband antennas that satisfy specific performance
criteria. Namely, the developed inverse design methodology enables
customization for the target impedance value, bandwidth, and contact
topology requirements. The proposed antenna achieves an improvement of
up to 40% in power transfer efficiency, compared with traditional
bow-tie antennas, under realistic operating conditions. High-fidelity
electromagnetic simulations validate these results, confirming the
design's practicality for terahertz applications. This work addresses
critical limitations of existing antenna designs and advances the
feasibility of high-frequency applications in both communication and
sensing.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Terahertz technology, a cornerstone of next-generation high-speed
communication and sensing, has long been hindered by impedance mismatch
challenges that limit device performance and applicability. These
challenges become particularly pronounced when ultrasensitive
two-dimensional (2D) materials are employed as the detecting element in
the terahertz range, further complicating their integration in realworld
applications. Furthermore, conventional antenna designs often fail to
provide adequate matching across the broad terahertz spectrum. In this
work, we tackle these challenges using a procedural generation algorithm
to design terahertz broadband antennas that satisfy specific performance
criteria. Namely, the developed inverse design methodology enables
customization for the target impedance value, bandwidth, and contact
topology requirements. The proposed antenna achieves an improvement of
up to 40% in power transfer efficiency, compared with traditional
bow-tie antennas, under realistic operating conditions. High-fidelity
electromagnetic simulations validate these results, confirming the
design's practicality for terahertz applications. This work addresses
critical limitations of existing antenna designs and advances the
feasibility of high-frequency applications in both communication and
sensing. - FNClarivate Analytics Web of Science
- VR1.0
- PTJ
- AFM Y Lukianov
A Maevskiy
N Kazeev
D Mylnikov
D A Svintsov
K S Novoselov
A Ustyuzhanin
D A Bandurin
- TIInverse design of broadband antennas for terahertz devices based on
two-dimensional materials - SOPHYSICAL REVIEW APPLIED
- DTArticle
- ABTerahertz technology, a cornerstone of next-generation high-speed
communication and sensing, has long been hindered by impedance mismatch
challenges that limit device performance and applicability. These
challenges become particularly pronounced when ultrasensitive
two-dimensional (2D) materials are employed as the detecting element in
the terahertz range, further complicating their integration in realworld
applications. Furthermore, conventional antenna designs often fail to
provide adequate matching across the broad terahertz spectrum. In this
work, we tackle these challenges using a procedural generation algorithm
to design terahertz broadband antennas that satisfy specific performance
criteria. Namely, the developed inverse design methodology enables
customization for the target impedance value, bandwidth, and contact
topology requirements. The proposed antenna achieves an improvement of
up to 40% in power transfer efficiency, compared with traditional
bow-tie antennas, under realistic operating conditions. High-fidelity
electromagnetic simulations validate these results, confirming the
design's practicality for terahertz applications. This work addresses
critical limitations of existing antenna designs and advances the
feasibility of high-frequency applications in both communication and
sensing. - Z90
- PUAMER PHYSICAL SOC
- PAONE PHYSICS ELLIPSE, COLLEGE PK, MD 20740-3844 USA
- SN2331-7019
- VL24
- DI10.1103/gr2z-3qjp
- UTWOS:001633477300003
- ER
- EF
|
Zhang, Pengxiang; Wang, Qian; Zhang, Yixin; Lin, Mo; Zhou, Xin; David, Ashish; Ustyuzhanin, Andrey; Chen, Musen; Katsnelson, Mikhail I; Trubyanov, Maxim; Novoselov, Kostya S; Andreeva, Daria V Strain-induced crumpling of graphene oxide lamellas to achieve fast and
selective transport of H2 and CO2 17 NATURE NANOTECHNOLOGY, 20 (9), pp. 1254-1261, 2025, DOI: 10.1038/s41565-025-01971-8. Abstract | BibTeX | Endnote @article{WOS:001528331800001,
title = {Strain-induced crumpling of graphene oxide lamellas to achieve fast and
selective transport of H2 and CO2},
author = {Pengxiang Zhang and Qian Wang and Yixin Zhang and Mo Lin and Xin Zhou and Ashish David and Andrey Ustyuzhanin and Musen Chen and Mikhail I Katsnelson and Maxim Trubyanov and Kostya S Novoselov and Daria V Andreeva},
doi = {10.1038/s41565-025-01971-8},
times_cited = {17},
issn = {1748-3387},
year = {2025},
date = {2025-09-01},
journal = {NATURE NANOTECHNOLOGY},
volume = {20},
number = {9},
pages = {1254-1261},
publisher = {NATURE PORTFOLIO},
address = {HEIDELBERGER PLATZ 3, BERLIN, 14197, GERMANY},
abstract = {Graphene oxide (GO) membranes offer high selectivity and
energy-efficient gas separation. However, their dense, layered structure
and tortuous diffusion paths limit permeability, posing a barrier to
industrial use. Here we present a method to enhance selectivity and
permeability, maintaining the structural stability of such membranes.
With an industrially friendly manufacturing method, we produce crumpled
GO membranes with gas diffusion pathways controlled by a multidomain
structure. These membranes achieve H2 permeability of approximately 2.1
x 104 barrer, significantly surpassing the permeability of flat lamellar
GO membranes, which is below 100 barrer. Its H2/CO2 selectivity of 91
outperforms current membrane technologies. In addition, the crumpled
membranes demonstrate stability under harsh conditions (-20 degrees C,
96% relative humidity), a critical requirement for practical
applications. This work addresses the long-standing
permeability-selectivity trade-off and establishes a robust, scalable
platform for integrating two-dimensional materials into membrane
technology for real-world applications.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Graphene oxide (GO) membranes offer high selectivity and
energy-efficient gas separation. However, their dense, layered structure
and tortuous diffusion paths limit permeability, posing a barrier to
industrial use. Here we present a method to enhance selectivity and
permeability, maintaining the structural stability of such membranes.
With an industrially friendly manufacturing method, we produce crumpled
GO membranes with gas diffusion pathways controlled by a multidomain
structure. These membranes achieve H2 permeability of approximately 2.1
x 104 barrer, significantly surpassing the permeability of flat lamellar
GO membranes, which is below 100 barrer. Its H2/CO2 selectivity of 91
outperforms current membrane technologies. In addition, the crumpled
membranes demonstrate stability under harsh conditions (-20 degrees C,
96% relative humidity), a critical requirement for practical
applications. This work addresses the long-standing
permeability-selectivity trade-off and establishes a robust, scalable
platform for integrating two-dimensional materials into membrane
technology for real-world applications. - FNClarivate Analytics Web of Science
- VR1.0
- PTJ
- AFPengxiang Zhang
Qian Wang
Yixin Zhang
Mo Lin
Xin Zhou
Ashish David
Andrey Ustyuzhanin
Musen Chen
Mikhail I Katsnelson
Maxim Trubyanov
Kostya S Novoselov
Daria V Andreeva
- TIStrain-induced crumpling of graphene oxide lamellas to achieve fast and
selective transport of H2 and CO2 - SONATURE NANOTECHNOLOGY
- DTArticle
- ABGraphene oxide (GO) membranes offer high selectivity and
energy-efficient gas separation. However, their dense, layered structure
and tortuous diffusion paths limit permeability, posing a barrier to
industrial use. Here we present a method to enhance selectivity and
permeability, maintaining the structural stability of such membranes.
With an industrially friendly manufacturing method, we produce crumpled
GO membranes with gas diffusion pathways controlled by a multidomain
structure. These membranes achieve H2 permeability of approximately 2.1
x 104 barrer, significantly surpassing the permeability of flat lamellar
GO membranes, which is below 100 barrer. Its H2/CO2 selectivity of 91
outperforms current membrane technologies. In addition, the crumpled
membranes demonstrate stability under harsh conditions (-20 degrees C,
96% relative humidity), a critical requirement for practical
applications. This work addresses the long-standing
permeability-selectivity trade-off and establishes a robust, scalable
platform for integrating two-dimensional materials into membrane
technology for real-world applications. - Z917
- PUNATURE PORTFOLIO
- PAHEIDELBERGER PLATZ 3, BERLIN, 14197, GERMANY
- SN1748-3387
- VL20
- BP1254
- EP1261
- DI10.1038/s41565-025-01971-8
- UTWOS:001528331800001
- ER
- EF
|
Rodin, A; Olsen, B A; Ustyuzhanin, A; Maevskiy, A Time-local stochastic equation of motion for solid ionic electrolytes PHYSICAL REVIEW RESEARCH, 7 (3), 2025, DOI: 10.1103/jnzr-q953. Abstract | BibTeX | Endnote @article{WOS:001547472300004,
title = {Time-local stochastic equation of motion for solid ionic electrolytes},
author = {A Rodin and B A Olsen and A Ustyuzhanin and A Maevskiy},
doi = {10.1103/jnzr-q953},
times_cited = {0},
year = {2025},
date = {2025-08-01},
journal = {PHYSICAL REVIEW RESEARCH},
volume = {7},
number = {3},
publisher = {AMER PHYSICAL SOC},
address = {ONE PHYSICS ELLIPSE, COLLEGE PK, MD 20740-3844 USA},
abstract = {Numerical studies of ionic motion through solid electrolytes commonly
involve static nudged-elastic band methods or costly ab initio molecular
dynamics. Building on a time-local model of current carrier-electrolyte
interaction and incorporating thermal motion, we introduce an approach
that is intermediate between the two well-established methodologies by
treating the electrolyte as an effective medium that interacts with the
mobile particle. Through this coupling, the thermally vibrating
electrolyte imparts energy to the charge carriers while also absorbing
energy from them due to its own finite elasticity. Using a simple model
system, we validate our approach through a series of numerical
simulations. Our methodology reproduces both dissipative and diffusive
behavior and helps link microscopic system parameters to measurable
macroscopic properties.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Numerical studies of ionic motion through solid electrolytes commonly
involve static nudged-elastic band methods or costly ab initio molecular
dynamics. Building on a time-local model of current carrier-electrolyte
interaction and incorporating thermal motion, we introduce an approach
that is intermediate between the two well-established methodologies by
treating the electrolyte as an effective medium that interacts with the
mobile particle. Through this coupling, the thermally vibrating
electrolyte imparts energy to the charge carriers while also absorbing
energy from them due to its own finite elasticity. Using a simple model
system, we validate our approach through a series of numerical
simulations. Our methodology reproduces both dissipative and diffusive
behavior and helps link microscopic system parameters to measurable
macroscopic properties. - FNClarivate Analytics Web of Science
- VR1.0
- PTJ
- AFA Rodin
B A Olsen
A Ustyuzhanin
A Maevskiy
- TITime-local stochastic equation of motion for solid ionic electrolytes
- SOPHYSICAL REVIEW RESEARCH
- DTArticle
- ABNumerical studies of ionic motion through solid electrolytes commonly
involve static nudged-elastic band methods or costly ab initio molecular
dynamics. Building on a time-local model of current carrier-electrolyte
interaction and incorporating thermal motion, we introduce an approach
that is intermediate between the two well-established methodologies by
treating the electrolyte as an effective medium that interacts with the
mobile particle. Through this coupling, the thermally vibrating
electrolyte imparts energy to the charge carriers while also absorbing
energy from them due to its own finite elasticity. Using a simple model
system, we validate our approach through a series of numerical
simulations. Our methodology reproduces both dissipative and diffusive
behavior and helps link microscopic system parameters to measurable
macroscopic properties. - Z90
- PUAMER PHYSICAL SOC
- PAONE PHYSICS ELLIPSE, COLLEGE PK, MD 20740-3844 USA
- VL7
- DI10.1103/jnzr-q953
- UTWOS:001547472300004
- ER
- EF
|
