-
The Bid Picture: Auction-Inspired Multi-player Generative Adversarial Networks Training
Authors:
Joo Yong Shim,
Jean Seong Bjorn Choe,
Jong-Kook Kim
Abstract:
This article proposes auction-inspired multi-player generative adversarial networks training, which mitigates the mode collapse problem of GANs. Mode collapse occurs when an over-fitted generator generates a limited range of samples, often concentrating on a small subset of the data distribution. Despite the restricted diversity of generated samples, the discriminator can still be deceived into di…
▽ More
This article proposes auction-inspired multi-player generative adversarial networks training, which mitigates the mode collapse problem of GANs. Mode collapse occurs when an over-fitted generator generates a limited range of samples, often concentrating on a small subset of the data distribution. Despite the restricted diversity of generated samples, the discriminator can still be deceived into distinguishing these samples as real samples from the actual distribution. In the absence of external standards, a model cannot recognize its failure during the training phase. We extend the two-player game of generative adversarial networks to the multi-player game. During the training, the values of each model are determined by the bids submitted by other players in an auction-like process.
△ Less
Submitted 20 March, 2024;
originally announced March 2024.
-
Designing Pr-based Advanced Photoluminescent Materials using Machine Learning and Density Functional Theory
Authors:
Upendra Kumar,
Hyeon Woo Kim,
Sobhit Singh,
Hyunseok Ko,
Sung Beom Cho
Abstract:
This work presents a machine learning approach to predict novel perovskite oxide materials in the Pr-Al-O and Pr-Sc-O compound families with the potential for photoluminescence applications. The predicted materials exhibit a large bandgap and high Debye temperature, and have remained unexplored thus far. The predicted compounds (Pr$_3$AlO$_6$, Pr$_4$Al$_2$O$_9$, Pr$_3$ScO$_6$ and Pr$_3$Sc$_5$O…
▽ More
This work presents a machine learning approach to predict novel perovskite oxide materials in the Pr-Al-O and Pr-Sc-O compound families with the potential for photoluminescence applications. The predicted materials exhibit a large bandgap and high Debye temperature, and have remained unexplored thus far. The predicted compounds (Pr$_3$AlO$_6$, Pr$_4$Al$_2$O$_9$, Pr$_3$ScO$_6$ and Pr$_3$Sc$_5$O$_{12}$) are screened using machine learning approach, which are then confirmed by density functional theory calculations. The study includes the calculation of the bandgap and density of states to determine electronic properties, and the optical absorption and emission spectra to determine optical properties. Mechanical stability of the predicted compounds, as demonstrated by satisfying the Born-Huang criterion. By combining machine learning and density functional theory, this work offers a more efficient and comprehensive approach to materials discovery and design.
△ Less
Submitted 20 June, 2023;
originally announced June 2023.
-
Highly tunable polarization-engineered two-dimensional electron gas in $ε$-AlGaO3 / $ε$-Ga2O3 heterostructures
Authors:
Praneeth Ranga,
Sung Beom Cho,
Rohan Mishra,
Sriram Krishnamoorthy
Abstract:
We report on the modeling of polarization-induced two-dimensional electron gas (2DEG) formation at $ε$-AlGaO3 / $ε$-Ga2O3 heterointerface and the effect of spontaneous polarization (Psp) reversal on 2DEG density in $ε$-Ga2O3 /$ε$-AlGaO3 / $ε$-Ga2O3 double heterostructures. Density-functional theory (DFT) is utilized to calculate the material properties of $ε$-Ga2O3 and $ε$-AlGaO3 alloys. Using Sch…
▽ More
We report on the modeling of polarization-induced two-dimensional electron gas (2DEG) formation at $ε$-AlGaO3 / $ε$-Ga2O3 heterointerface and the effect of spontaneous polarization (Psp) reversal on 2DEG density in $ε$-Ga2O3 /$ε$-AlGaO3 / $ε$-Ga2O3 double heterostructures. Density-functional theory (DFT) is utilized to calculate the material properties of $ε$-Ga2O3 and $ε$-AlGaO3 alloys. Using Schrodinger-Poisson solver along with DFT calculated parameters, the 2DEG density is calculated as a function of barrier type and thickness. By optimizing the layer thicknesses of $ε$-Ga2O3/$ε$-AlGaO3/$ε$-Ga2O3 heterostructures, charge contrast ratios exceeding 1600 are obtained. This computational study indicates the high potential for $ε$-Ga2O3-based heterostructure devices for non-volatile memories and neuromorphic applications.
△ Less
Submitted 28 April, 2020;
originally announced April 2020.
-
Spin-polarized bandgap of graphene induced by alternative chemisorption with MgO (111) substrate
Authors:
Sung Beom Cho,
Yong-Chae Chung
Abstract:
Using First-principle calculations, substrate effect of O-terminated (rt3 x rt3) MgO (111) on graphene was investigated for spintronics application. Surprisingly, the graphene can be turned into a spin-polarized semiconductor, which implies that the totally spin-polarized current can be generated and its on/off switching can be also controlled. The origin of the spin-polarized band structure is sp…
▽ More
Using First-principle calculations, substrate effect of O-terminated (rt3 x rt3) MgO (111) on graphene was investigated for spintronics application. Surprisingly, the graphene can be turned into a spin-polarized semiconductor, which implies that the totally spin-polarized current can be generated and its on/off switching can be also controlled. The origin of the spin-polarized band structure is spin-ordering due to alternative sp2-sp3 covalent bondings induced by the MgO (111) substrate. The results indicate that the tailored pattern of the chemisorption can be highly efficient or introducing totally spin-polarized current to the graphene.
△ Less
Submitted 25 November, 2013; v1 submitted 13 May, 2013;
originally announced May 2013.
