Search for: All records

Face recognition systems have made significant strides thanks to data-heavy deep learning models, but these models rely on large privacy-sensitive datasets. Recent work in facial analysis and recognition have thus started making use of synthetic datasets generated from GANs and diffusion based generative models. These models, however, lack fairness in terms of demographic representation and can introduce the same biases in the trained downstream tasks. This can have serious societal and security implications. To address this issue, we propose a methodology that generates unbiased data from a biased generative model using an evolutionary algorithm. We show results for StyleGAN2 model trained on the Flicker Faces High Quality dataset to generate data for singular and combinations of demographic attributes such as Black and Woman. We generate a large racially balanced dataset of 13.5 million images, and show that it boosts the performance of facial recognition and analysis systems whilst reducing their biases. We have made our code-base ( https://github.com/anubhav1997/youneednodataset ) public to allow researchers to reproduce our work. 

A dictionary attack in a biometric system entails the use of a small number of strategically generated images or tem- plates to successfully match with a large number of identi- ties, thereby compromising security. We focus on dictionary attacks at the template level, specifically the IrisCodes used in iris recognition systems. We present an hitherto unknown vulnerability wherein we mix IrisCodes using simple bit- wise operators to generate alpha-mixtures —alpha-wolves (combining a set of “wolf” samples) and alpha-mammals (combining a set of users selected via search optimization) that increase false matches. We evaluate this vulnerabil- ity using the IITD, CASIA-IrisV4-Thousand and Synthetic datasets, and observe that an alpha-wolf (from two wolves) can match upto 71 identities @FMR=0.001%, while an alpha-mammal (from two identities) can match upto 133 other identities @FMR=0.01% on the IITD dataset. 

Abstract Deep learning (DL) is one of the fastest-growing topics in materials data science, with rapidly emerging applications spanning atomistic, image-based, spectral, and textual data modalities. DL allows analysis of unstructured data and automated identification of features. The recent development of large materials databases has fueled the application of DL methods in atomistic prediction in particular. In contrast, advances in image and spectral data have largely leveraged synthetic data enabled by high-quality forward models as well as by generative unsupervised DL methods. In this article, we present a high-level overview of deep learning methods followed by a detailed discussion of recent developments of deep learning in atomistic simulation, materials imaging, spectral analysis, and natural language processing. For each modality we discuss applications involving both theoretical and experimental data, typical modeling approaches with their strengths and limitations, and relevant publicly available software and datasets. We conclude the review with a discussion of recent cross-cutting work related to uncertainty quantification in this field and a brief perspective on limitations, challenges, and potential growth areas for DL methods in materials science. 

Computational prediction of good thermoelectric (TE) performance in several n-type doped Zintl phases, combined with successful experimental realization, has sparked interest in discovering new n-type dopable members of this family of materials. However, most known Zintls are typically only p-type dopable; prior successes in finding n-type Zintl phases have been largely serendipitous. Here, we go beyond previously synthesized Zintl phases and perform chemical substitutions in known n-type dopable ABX Zintl phases to discover new ones. We use first-principles calculations to predict their stability, potential for TE performance as well as their n-type dopability. Using this approach, we find 17 new ABX Zintl phases in the KSnSb structure type that are predicted to be stable. Several of these newly predicted phases (KSnBi, RbSnBi, NaGeP) are found to exhibit promising n-type TE performance and are n-type dopable. We propose these compounds for further experimental studies, especially KSnBi and RbSnBi, which are both predicted to be good TE materials with high electron concentrations due to self-doping by native defects, when grown under alkali-rich conditions. 

Abstract This Roadmap article provides a succinct, comprehensive overview of the state of electronic structure (ES) methods and software for molecular and materials simulations. Seventeen distinct sections collect insights by 51 leading scientists in the field. Each contribution addresses the status of a particular area, as well as current challenges and anticipated future advances, with a particular eye towards software related aspects and providing key references for further reading. Foundational sections cover density functional theory and its implementation in real-world simulation frameworks, Green’s function based many-body perturbation theory, wave-function based and stochastic ES approaches, relativistic effects and semiempirical ES theory approaches. Subsequent sections cover nuclear quantum effects, real-time propagation of the ES, challenges for computational spectroscopy simulations, and exploration of complex potential energy surfaces. The final sections summarize practical aspects, including computational workflows for complex simulation tasks, the impact of current and future high-performance computing architectures, software engineering practices, education and training to maintain and broaden the community, as well as the status of and needs for ES based modeling from the vantage point of industry environments. Overall, the field of ES software and method development continues to unlock immense opportunities for future scientific discovery, based on the growing ability of computations to reveal complex phenomena, processes and properties that are determined by the make-up of matter at the atomic scale, with high precision. 

Half-Heusler materials are strong candidates for thermoelectric applications due to their high weighted mobilities and power factors, which is known to be correlated to valley degeneracy in the electronic band structure. However, there are over 50 known semiconducting half-Heusler phases, and it is not clear how the chemical composition affects the electronic structure. While all the n-type electronic structures have their conduction band minimum at either the Γ - or X -point, there is more diversity in the p-type electronic structures, and the valence band maximum can be at either the Γ -, L -, or W -point. Here, we use high throughput computation and machine learning to compare the valence bands of known half-Heusler compounds and discover new chemical guidelines for promoting the highly degenerate W -point to the valence band maximum. We do this by constructing an “orbital phase diagram” to cluster the variety of electronic structures expressed by these phases into groups, based on the atomic orbitals that contribute most to their valence bands. Then, with the aid of machine learning, we develop new chemical rules that predict the location of the valence band maximum in each of the phases. These rules can be used to engineer band structures with band convergence and high valley degeneracy. 

Accurate density functional theory calculations of the interrelated properties of thermoelectric materials entail high computational cost, especially as crystal structures increase in complexity and size. New methods involving ab initio scattering and transport (AMSET) and compressive sensing lattice dynamics are used to compute the transport properties of quaternary CaAl 2 Si 2 -type rare-earth phosphides RECuZnP 2 (RE = Pr, Nd, Er), which were identified to be promising thermoelectrics from high-throughput screening of 20 000 disordered compounds. Experimental measurements of the transport properties agree well with the computed values. Compounds with stiff bulk moduli (>80 GPa) and high speeds of sound (>3500 m s −1 ) such as RECuZnP 2 are typically dismissed as thermoelectric materials because they are expected to exhibit high lattice thermal conductivity. However, RECuZnP 2 exhibits not only low electrical resistivity, but also low lattice thermal conductivity (∼1 W m −1 K −1 ). Contrary to prior assumptions, polar-optical phonon scattering was revealed by AMSET to be the primary mechanism limiting the electronic mobility of these compounds, raising questions about existing assumptions of scattering mechanisms in this class of thermoelectric materials. The resulting thermoelectric performance ( zT of 0.5 for ErCuZnP 2 at 800 K) is among the best observed in phosphides and can likely be improved with further optimization. 

Engineering the thermal properties in solids is important for both fundamental physics ( e.g. electric and phonon transport) and device applications ( e.g. thermal insulating coating, thermoelectrics). In this paper, we report low thermal transport properties of four selenide compounds (BaAg 2 SnSe 4 , BaCu 2 GeSe 4 , BaCu 2 SnSe 4 and SrCu 2 GeSe 4 ) with experimentally-measured thermal conductivity as low as 0.31 ± 0.03 W m −1 K −1 at 673 K for BaAg 2 SnSe 4 . Density functional theory calculations predict κ < 0.3 W m −1 K −1 for BaAg 2 SnSe 4 due to scattering from weakly-bonded Ag–Ag dimers. Defect calculations suggest that achieving high hole doping levels in these materials could be challenging due to monovalent ( e.g. , Ag) interstitials acting as hole killers, resulting in overall low electrical conductivity in these compounds.