RSS 1.0Articles
Recent Submissions
-
Quantifying Within-Flight Variation in Land Surface Temperature from a UAV-Based Thermal Infrared Camera(Drones, MDPI AG, 2023-10-02) [Article]Land Surface Temperature (LST) is a key variable used across various applications, including irrigation monitoring, vegetation health assessment and urban heat island studies. While satellites offer moderate-resolution LST data, unmanned aerial vehicles (UAVs) provide high-resolution thermal infrared measurements. However, the continuous and rapid variation in LST makes the production of orthomosaics from UAV-based image collections challenging. Understanding the environmental and meteorological factors that amplify this variation is necessary to select the most suitable conditions for collecting UAV-based thermal data. Here, we capture variations in LST while hovering for 15–20 min over diverse surfaces, covering sand, water, grass, and an olive tree orchard. The impact of different flying heights and times of the day was examined, with all collected thermal data evaluated against calibrated field-based Apogee SI-111 sensors. The evaluation showed a significant error in UAV-based data associated with wind speed, which increased the bias from −1.02 to 3.86 °C for 0.8 to 8.5 m/s winds, respectively. Different surfaces, albeit under varying ambient conditions, showed temperature variations ranging from 1.4 to 6 °C during the flights. The temperature variations observed while hovering were linked to solar radiation, specifically radiation fluctuations occurring after sunrise and before sunset. Irrigation and atmospheric conditions (i.e., thin clouds) also contributed to observed temperature variations. This research offers valuable insights into LST variations during standard 15–20 min UAV flights under diverse environmental conditions. Understanding these factors is essential for developing correction procedures and considering data inconsistencies when processing and interpreting UAV-based thermal infrared data and derived orthomosaics.
-
A Wideband Transition Design Technique from RWG to SIW Technologies(IEEE Access, Institute of Electrical and Electronics Engineers (IEEE), 2023-10-02) [Article]Efficient wide-band transitions play a critical role in integrating substrate-integrated waveguide (SIW) and air-filled substrate-integrated waveguide (AFSIW) systems with conventional rectangular waveguide (RWG) systems in millimeter wave communication. However, existing techniques lack a systematic design approach, overlook mode purity, encounter performance issues, involve tedious optimizations, and feature complex structures. This paper presents a first-of-its-kind design technique that addresses these limitations, enabling seamless transitions from RWG to both SIW and AFSIW. For the first time, a clear and systematic design flow is introduced, covering both transition scenarios. The proposed technique eliminates the need for separate design procedures, simplifying the process, reducing complexity, and offering a cost-effective solution with state-of-the-art performance. The key innovation lies in the analytical expression for the shortest transition length, derived for both SIW and AFSIW cases, facilitating the design process and optimizing transition performance. The effectiveness of the design flow is extensively validated through four design examples, evaluating mode purity, reflection coefficient, and insertion loss. The results demonstrate excellent performances, confirming the accuracy of the approach. Furthermore, a fabricated prototype achieves remarkable results with an insertion loss of 0.35 dB and a relative bandwidth of 40%, exhibiting state-of-the-art performances when compared to reported works in the literature. In summary, the novel design technique enables seamless wide-band transitions from RWG to both SIW and AFSIW, offering a unified approach with state-of-the-art performance.
-
Face-directed assembly of tailored isoreticular MOFs using centring structure-directing agents(Nature Synthesis, Springer Science and Business Media LLC, 2023-10-02) [Article]Building blocks with low connectivity and no embedded directionality are prone to polymorphism, as demonstrated by the diversity of 4-connected zeolitic nets (>250). As a result, their deployment for design in reticular and isoreticular chemistries remains a challenge. However, the ability to control geometrical peculiarities offers potential to deviate from the assembly of default structures. Here we report the face-directed assembly of >20 isoreticular zeolite-like metal–organic frameworks (ZMOFs) by using polytopic expanding and tightening centring structure-directing agents (cSDAs). The cSDAs are selected with the appropriate geometrical coding information to alter and control the orientation of adjacent supermolecular building blocks. The ZMOFs have an underlying sodalite (sod) topology that is remarkably suited for the rational assembly of multinary materials. In addition to a variety of metal cations (In, Fe, Co and Ni), a diverse range of cSDAs (di-, tri-, tetra-, hexa-, pyridyl or imidazole) are used and combined. Our approach enables isoreticular possibilities at both extremities of the porous materials spectrum: In-sod-ZMOF-102 exhibits small pore aperture suitable for efficient separation, while Fe-sod-ZMOF-320 with 48-Å-wide mesopores exhibits high hydrogen uptake, methane storage working capacity and a high gravimetric working capacity for oxygen.
-
Pt–Fe Nanoparticles Dispersed on Mesoporous Silica as Selective Catalysts for Dehydrogenation of Isobutane(ACS Applied Nano Materials, American Chemical Society (ACS), 2023-10-02) [Article]A series of catalysts composed of Pt–Fe nanoparticles supported on mesoporous silica SBA-15 (ca. 7 nm pore diameter) have been prepared by ultrasound-assisted coimpregnation of the metal precursors and evaluated in the nonoxidative dehydrogenation of isobutane. Prereduced catalytic systems were characterized by STEM-HAADF + EDS mapping and XAS to determine the chemical environment of the highly dispersed platinum active sites on the iron host matrix. While the monometallic platinum (nanoparticles) supported on SBA-15 material presented a rapid catalyst deactivation under reaction conditions, coordinatively unsaturated Pt–Fe (Pt ≪ Fe) sites located in the mesopores of SBA-15 showed a high steady-state activity (43% conversion) and selectivity (96% to isobutylene) in the dehydrogenation of isobutane at 550 °C for several hours. Temperature-programmed reduction profiles determined not only the substantially higher reducibility of FeOx species with doping amounts of platinum in the as-prepared (calcined) catalysts but also the detrimental structural changes undergone after consecutive reaction–regeneration cycles. Finally, reactivation under controlled conditions allows to minimize irreversible catalyst deactivation after successive cycles.
-
A Tunable Micromachined Multithreshold Inertial Switch(IEEE/ASME Transactions on Mechatronics, Institute of Electrical and Electronics Engineers (IEEE), 2023-09-29) [Article]In this article, we present a multithreshold microelectromechanical tunable inertial switch. The device aims to provide quantitative information on acceleration while retaining the attractive energy-saving features of binary threshold switches. The designed proof-of-concept device with three thresholds is composed of four serpentine springs, a suspended proof mass, and three stationary electrodes placed at various positions in the sensing direction. In addition, the tunability of the acceleration threshold is demonstrated based on the softening effect of the electrostatic force. The dynamic behavior of the switch is investigated analytically. Results are shown for the relationship between the bias voltage and tunable threshold. The fabricated switch prototypes are tested using a drop-table shock system. The test results demonstrate that the multithreshold switch can detect an acceleration range of 131–400 g. The simulated and analytical results are in good agreement with the experimental data. The demonstrated device concept is promising to categorize the shock impact for several applications, such as for head impact and brain injuries.
-
Design Criteria for Silicon Solar Cells with Fill Factors Approaching the Auger Limit(ACS Energy Letters, American Chemical Society (ACS), 2023-09-29) [Article]We establish, via a systematic simulation study, the minimum requirements for the electrical design parameters to accomplish fill factors above 86% in crystalline-silicon solar cells.
-
Developing a data-driven modeling framework for simulating a chemical accident in freshwater(Journal of Cleaner Production, Elsevier BV, 2023-09-29) [Article]Chemical accidents in freshwater pose threats to public health and aquatic ecosystems. Process-based models (PBMs) have been used to identify spatiotemporal chemical distributions in natural water. However, their computationally expensive simulations can hinder timely incident responses, which are crucial for minimizing negative impacts. Therefore, this study proposes a site-specific data-driven model (DDM) to supplement PBM-based chemical accident simulations. A convolutional neural network (CNN) was employed as the DDM because of its outstanding performance in capturing spatial patterns. Our model was developed to facilitate chemical accident simulations in the Namhan River, South Korea. The model datasets were generated using the PBM simulation outputs from toluene accident scenarios. Our DDM showed a Nash-Sutcliffe-efficiency of 0.94 and a root-mean-square-error of 0.023 μg/L for the validation set. Its computational time was approximately 64 times faster than that of PBMs. In addition, this study interpreted the DDM results using SHapley Additive exPlanations (SHAP). The SHAP findings highlighted the influential role of distance from the accident site in this study. Overall, this study demonstrated the applicability of our modeling approach in freshwater chemical accidents by providing rapid spatial distribution results complementing PBM simulations.
-
Development and Validation of the Workplace Affective Events Survey(Cureus, Springer Science and Business Media LLC, 2023-09-29) [Article]Background Work, a central aspect of human life, serves vital economic and social functions. There is a burgeoning interest in positive emotions in the workplace, which can enhance creativity, foster social connections, and improve problem-solving abilities. These emotions are pivotal in key organizational outcomes, including employee performance and health. Despite the extensive examination of factors like job satisfaction and workplace stressors, a knowledge gap exists regarding the everyday workplace events that influence emotions and their contribution to overall workplace emotional health. The present study introduces the Workplace Affective Events Survey (WAES), a new tool that can facilitate the advancement of research in this field. Purpose This study aimed to develop a tool to assess daily workplace events that lead to positive or negative emotional responses and the intensities of such responses. The study also examined the relationship between these events and the associated affect-intensities with trait affect, and social companionship at work for convergent validation. Methodology The tool development entailed a multi-phase approach which encompassed item generation, content validation, pre-pilot trials, and pilot testing of the WAES. Participants were entry and mid-level service sector employees aged 25-55 years. Themes generated using focus group discussions and one-to-one interviews were mapped against a known taxonomy of workplace affective events. Expert validation and pre-pilot trials helped in refining the final items. The main phase engaged 300 individuals from nine service industries across 29 organizations in an urban metropolitan city in India. WAES was administered alongside standardized measures of trait-affect and workplace social companionship. Results WAES subscales demonstrated acceptable reliability. Participants reported positive daily affective events more often than negative ones, with the average intensity of positive emotions surpassing that of negative emotions. Notably, trait affect scores and social companionship exhibited significant correlations with daily affective events and their intensity. Conclusions The WAES offers a novel tool to investigate daily emotional experiences in the workplace. The data suggest that a within-person disposition such as trait-affect might play a lesser role in generating positive affective events than contextual factors. These findings underscore the value of creating work environments that consistently nurture positive emotional experiences.
-
Chemiluminescence- and machine learning-based monitoring of premixed ammonia-methane-air flames(Applications in Energy and Combustion Science, Elsevier BV, 2023-09-29) [Article]This work presents the development and validation of an algorithm capable of predicting the equivalence ratio and the ammonia fraction of premixed ammonia-methane-air flames using only measured OH*, NH*, CN*, and CH* chemiluminescence intensities as input. This machine learning algorithm relies on Gaussian process regression (GPR). It was trained and validated with data previously recorded in laminar flames, and it was then tested with new data recorded in more practical, turbulent swirl flames. The algorithm performs well for laminar and turbulent flames for wide ranges of equivalence ratio (0.80 ≤ ϕ ≤ 1.20) and ammonia fraction (0 ≤ XNH3 ≤ 0.60). For turbulent swirl flames, the prediction errors in the equivalence ratio and on the ammonia fraction are smaller than 0.05, except for a very small subset of operating conditions where the error is up to 0.10. Additional tests were performed by adding NO* and CO2* to the list of inputs, but this did not improve the predictions. The GPR algorithm was then benchmarked against linear and polynomial regressions and a more conventional way of inferring flame properties from chemiluminescence measurements, namely the ratio-based method. This method relies only on CN*/NO* and NH*/CH* ratios to predict the equivalence ratio and the ammonia fraction. Its prediction errors were often larger than 0.15, which is significantly worse than that of the GPR algorithm. Consequently, this work constitutes a solid basis for the future development of non-intrusive sensors to monitor practical ammonia-methane-air flames.
-
A strategy For High Ethylene Polymerization Performance Using Titanium Single-Site Catalysts(Chemical Communications, Royal Society of Chemistry (RSC), 2023-09-29) [Article]The synthesis of heterogeneous Ti(IV)-based catalysts for ethylene polymerization following surface organometallic chemistry concepts is described. The unique feature of this catalyst arises from the silica support, KCC-1700. It has (i) a 3D fibrous morphology that is essential to improve the diffusion of the reactants, and (ii) an aluminum-bound hydroxyl group, [([triple bond, length as m-dash]Si–O–Si[triple bond, length as m-dash])([triple bond, length as m-dash]Si–O–)2Al–OH] 2, used as an anchoring site. The [([triple bond, length as m-dash]Si–O–Si[triple bond, length as m-dash])([triple bond, length as m-dash]Si–O–)(Al–O–)TiNp3] 3 catalyst was obtained by reacting 2 with a tetrakis-(neopentyl) titanium TiNp4. The structure of 3 was fully characterized by FT-IR, advanced solid-state NMR spectroscopy [1H, 13C], elemental and gas-phase analysis (ICP-OES and CHNS analysis), and XPS. The benefits of combining these morphological (3D structure) and electronic properties of the support (aluminum plus titanium) were evidenced in ethylene polymerization. The results show a remarkable enhancement in the catalytic performance with the formation of HDPE. Notably, the resulting HDPE displays a molecular weight of 3 200 000 g mol−1 associated with a polydispersity index (PD) of 2.3. Moreover, the effect of the mesostructure (2D vs. 3D) was demonstrated in the catalytic activity for ethylene polymerization.
-
Transgenerational adaptation to ocean acidification determines the susceptibility of filter-feeding rotifers to nanoplastics.(Journal of hazardous materials, Elsevier BV, 2023-09-28) [Article]The adaptation of marine organisms to the impending challenges presented by ocean acidification (OA) is essential for their future survival, and mechanisms underlying OA adaptation have been reported in several marine organisms. In the natural environment, however, marine organisms are often exposed to a combination of environmental stressors, and the interactions between adaptive responses have yet to be elucidated. Here, we investigated the susceptibility of filter-feeding rotifers to short-term (ST) and long-term (LT) (≥180 generations) high CO2 conditions coupled with nanoplastic (NPs) exposure (ST+ and LT+). Adaptation of rotifers to elevated CO2 caused differences in ingestion and accumulation of NPs, resulting in a significantly different mode of action on in vivo endpoints between the ST+ and LT+ groups. Moreover, microRNA-mediated epigenetic regulation was strongly correlated with the varied adaptive responses between the ST+ and LT+ groups, revealing novel regulatory targets and pathways. Our results indicate that pre-exposure history to increased CO2 levels is an important factor in the susceptibility of rotifers to NPs.
-
Strigolactone biosynthesis lgs1 mutant alleles mined from the sorghum accession panel are a promising resource of resistance to witchweed (Striga) parasitism(PLANTS, PEOPLE, PLANET, Wiley, 2023-09-28) [Article]Striga is a parasitic plant that greatly limits the production of Africa's most staple cereals, including sorghum. Infection occurs when the parasite germinates in response to biomolecules emitted into the soil from the host's roots. Some sorghum genotypes harbor a mutation that makes them ineffective in stimulating Striga seed germination. This resistance is of great importance because of its possible application in Striga management. Here, additional resistant sorghum genotypes with varying levels of Striga resistance are discussed in the context of their candidacy for integration in breeding programs and their possible role in alleviating food insecurity in sub-Saharan Africa by reducing crop losses because of Striga infestation. Sorghum is a food staple for millions of people in sub-Saharan Africa, but its production is greatly diminished by Striga, a parasitic weed. An efficient and cost-effective way of managing Striga in smallholder farms in Africa is to deploy resistant varieties of sorghum. Here, we leverage genomics and the vast genetic diversity of sorghum—evolutionarily adapted to cope with Striga parasitism in Africa—to identify new Striga-resistant sorghum genotypes by exploiting a resistance mechanism hinged on communication molecules called strigolactones (SLs), exuded by hosts to trigger parasite seed germination. We achieved this by mining for mutant alleles of the LOW GERMINATION STIMULANT 1 (LGS1) that are ineffective in stimulating Striga germination from the sorghum accession panel (SAP). Our analysis identified lgs1 sorghum genotypes, which we named SAP-lgs1. SAP-lgs1 had the SL exudation profile of known lgs1 sorghum, whose hallmark is the production of the low inducer of germination, orobanchol. Laboratory and field resistance screens showed that the SAP-lgs1 genotypes also exhibited remarkable resistance against Striga. Our findings have the potential to reduce crop losses because of Striga parasitism and therefore have far-reaching implications for improving food security in Africa
-
Dipeptide-Based Photoreactive Instant Glue for Environmental and Biomedical Applications.(ACS applied materials & interfaces, American Chemical Society (ACS), 2023-09-28) [Article]Nature-inspired smart materials offer numerous advantages over environmental friendliness and efficiency. Emulating the excellent adhesive properties of mussels foot proteins, where the lysine is in close proximity with the 3,4-dihydroxy-l-phenylalanine (DOPA), we report the synthesis of a novel photocurable peptide-based adhesive consisting exclusively of these two amino acids. Our adhesive is a highly concentrated aqueous solution of a monomer, a cross-linker, and a photoinitiator. Lap-shear adhesion measurements on plastic and glass surfaces and comparison with different types of commercial adhesives showed that the adhesive strength of our glue is comparable when applied in air and superior when used underwater. No toxicity of our adhesive was observed when the cytocompatibility on human dermal fibroblast cells was assessed. Preliminary experiments with various tissues and coral fragments showed that our adhesive could be applied to wound healing and coral reef restoration. Given the convenience of the facile synthesis, biocompatibility, ease of application underwater, and high adhesive strength, we expect that our adhesive may find application, but not limited, to the biomedical and environmental field.
-
On the Peak AoI of UAV-Assisted IoT Networks: A Stochastic Geometry Approach(IEEE Internet of Things Journal, Institute of Electrical and Electronics Engineers (IEEE), 2023-09-28) [Article]In this paper, we analyze the peak age of information (PAoI) in UAV-assisted internet of thing (IoT) networks, in which the locations of IoT devices are modeled by a Matérn cluster process (MCP) and UAVs are deployed at the cluster centers to collect the status updates from the devices. Specifically, we consider that IoT devices can either monitor the same physical process or different physical processes and UAVs split their resources, time or bandwidth, to serve the devices to avoid inter-cluster interference. Using tools from stochastic geometry, we are able to compute the mean activity probability of IoT devices and the conditional success probability of an individual device. We then use tools from queuing theory to compute the PAoI under two load models and two scenarios for devices, respectively. Our numerical results show interesting system insights. We first show that for a low data arrival rate, increasing the number of correlated devices can improve the PAoI for both load models. Next, we show that even though the time-splitting technique causes higher interference, it has a limited impact on the mean PAoI, and the mean PAoI benefits more from the time-splitting technique. This is because of the nature of UAV communication, especially at places where devices (users) are spatially-clustered: shorter transmission distances and better communication channels, comparing the links established by the cluster UAV and serving devices (users) to links established by interferers.
-
Enhanced optoelectronic coupling for perovskite-silicon tandem solar cells(Nature, Springer Science and Business Media LLC, 2023-09-28) [Article]Monolithic perovskite/silicon tandem solar cells are of great appeal as they promise high power conversion efficiencies (PCEs) at affordable cost. In state-of-the-art tandems, the perovskite top cell is electrically coupled to a silicon heterojunction bottom cell via a self-assembled monolayer (SAM), anchored on a transparent conductive oxide (TCO), which enables efficient charge transfer between the subcells.1-3 Yet, reproducible high-performance tandem solar cells require energetically homogenous SAM coverage, which remains challenging, especially on textured silicon bottom cells. Here, we resolve this issue by employing ultrathin (5 nm) amorphous indium zinc oxide (IZO) as the interconnecting TCO, exploiting its high surface-potential homogeneity resulting from the absence of crystal grains, and higher density of SAM anchoring sites when compared to commonly employed crystalline TCOs. Combined with optical enhancements via equally thin IZO rear electrodes and improved front contact stacks, an independently certified PCE of 32.5% was obtained, which ranks amongst the highest for perovskite/silicon tandems. Our ultrathin transparent contact approach reduces indium consumption by approximately 80%, which is of importance towards sustainable PV manufacturing.
-
High potential for biomass-degrading CAZymes revealed by pine forest soil metagenomics(Journal of Biomolecular Structure and Dynamics, Informa UK Limited, 2023-09-28) [Article]The undisturbed environment in Netarhat, with its high levels of accumulated lignocellulosic biomass, presents an opportunity to identify microbes for biomass digestion. This study focuses on the bioprospecting of native soil microbes from the Netarhat forest in Jharkhand, India, with the potential for lignocellulosic substrate digestion. These biocatalysts could help overcome the bottleneck of biomass saccharification and reduce the overall cost of biofuel production, replacing harmful fossil fuels. The study used metagenomic analysis of pine forest soil via whole genome shotgun sequencing, revealing that most of the reads matched with the bacterial species, very low percentage of reads (0.1%) belongs to fungal species, with 13% of unclassified reads. Actinobacteria were found to be predominant among the bacterial species. MetaErg annotation identified 11,830 protein family genes and 2 metabolic marker genes in the soil samples. Based on the Carbohydrate Active EnZyme (CAZy) database, 3,996 carbohydrate enzyme families were identified, with family Glycosyl hydrolase (GH) dominating with 1,704 genes. Most observed GH families in the study were GH0, 3, 5, 6. 9, 12. 13, 15, 16, 39, 43, 57, and 97. Modelling analysis of a representative GH 43 gene suggested a strong affinity for cellulose than xylan. This study highlights the lignocellulosic digestion potential of the native microfauna of the lesser-known pine forest of Netarhat.
-
Hybrid Multiple-Access: Mode Selection, User Pairing and Resource Allocation(IEEE Access, Institute of Electrical and Electronics Engineers (IEEE), 2023-09-28) [Article]This paper proposes hybridization of non-orthogonal multiple-access (NOMA) and orthogonal multiple access (OMA) schemes for next-generation cellular networks. Two hybrid schemes that operate NOMA/OMA mode selection as well as channel and power allocation are proposed to improve the resource utilization and bandwidth-efficiency for network capacity maximization. The two proposed hybrid schemes are: (1) single-cell hybrid multiple-access (SC-HMA) and (2) multi-cell hybrid multiple-access (MC-HMA) schemes. The SC-HMA scheme determines the optimal NOMA/OMA mode and user pairs in each resource block by utilizing a capacity matrix representing the capacity outcomes of pairing all possible combinations of users. On the other hand, in the MC-HMA, the NOMA/OMA modes are categorized into intra-cell and inter-cell based on an interference map, where the principal objective is to determine the best mode of operation between the user pairs to improve the overall sum-rate and quality-of-service (QoS). The results show that the proposed HMA schemes provide superior overall network capacity compared to the benchmark schemes. In addition, the SC-HMA scheme outperforms the MC-HMA in terms of network capacity at the expense of higher computational complexity.
-
Finite Sample Guarantees of Differentially Private Expectation Maximization Algorithm(Frontiers in Artificial Intelligence and Applications, IOS Press, 2023-09-28) [Article](Gradient) Expectation Maximization (EM) is a widely used algorithm for estimating the maximum likelihood of mixture models or incomplete data problems. A major challenge facing this popular technique is how to effectively preserve the privacy of sensitive data. Previous research on this problem has already lead to the discovery of some Differentially Private (DP) algorithms for (Gradient) EM. However, unlike in the non-private case, existing techniques are not yet able to provide finite sample statistical guarantees. To address this issue, we propose in this paper the first DP version of Gradient EM algorithm with statistical guarantees. Specifically, we first propose a new mechanism for privately estimating the mean of a heavy-tailed distribution, which significantly improves a previous result in [25], and it could be extended to the local DP model, which has not been studied before. Next, we apply our general framework to three canonical models: Gaussian Mixture Model (GMM), Mixture of Regressions Model (MRM) and Linear Regression with Missing Covariates (RMC). Specifically, for GMM in the DP model, our estimation error is near optimal in some cases. For the other two models, we provide the first result on finite sample statistical guarantees. Our theory is supported by thorough numerical experiments on both real-world data and synthetic data.
-
Wafer-Scale Memristor Array Based on Aligned Grain Boundaries of 2D Molybdenum Ditelluride for Application to Artificial Synapses(Advanced Functional Materials, Wiley, 2023-09-27) [Article]2D materials have attracted attention in the field of neuromorphic computing applications, demonstrating the potential for their use in low-power synaptic devices at the atomic scale. However, synthetic 2D materials contain randomly distributed intrinsic defects and exhibit a stochasitc forming process, which results in variability of switching voltages, times, and stat resistances, as well as poor synaptic plasticity. Here, this work reports the wafer-scale synthesis of highly polycrystalline semiconducting 2H-phase molybdenum ditelluride (2H-MoTe2) and its use for fabricating crossbar arrays of memristors. The 2H-MoTe2 films contain small grains (≈30 nm) separated by vertically aligned grain boundaries (GBs). These aligned GBs provide confined diffusion paths for metal ions filtration (from the electrodes), resulting in reliable resistive switching (RS) due to conductive filament confinement. As a result, the polycrystalline 2H-MoTe2 memristors shows improvement in the RS uniformity and stable multilevel resistance states, small cycle-to-cycle variation (<8.3%), high yield (>83.7%), and long retention times (>104 s). Finally, 2H-MoTe2 memristors show linear analog synaptic plasticity under more than 2500 repeatable pulses and a simulation-based learning accuracy of 96.05% for image classification, which is the first analog synapse behavior reported for 2D MoTe2 based memristors.
-
Unified Analysis of Stochastic Gradient Methods for Composite Convex and Smooth Optimization(Journal of Optimization Theory and Applications, Springer Science and Business Media LLC, 2023-09-27) [Article]We present a unified theorem for the convergence analysis of stochastic gradient algorithms for minimizing a smooth and convex loss plus a convex regularizer. We do this by extending the unified analysis of Gorbunov et al. (in: AISTATS, 2020) and dropping the requirement that the loss function be strongly convex. Instead, we rely only on convexity of the loss function. Our unified analysis applies to a host of existing algorithms such as proximal SGD, variance reduced methods, quantization and some coordinate descent-type methods. For the variance reduced methods, we recover the best known convergence rates as special cases. For proximal SGD, the quantization and coordinate-type methods, we uncover new state-of-the-art convergence rates. Our analysis also includes any form of sampling or minibatching. As such, we are able to determine the minibatch size that optimizes the total complexity of variance reduced methods. We showcase this by obtaining a simple formula for the optimal minibatch size of two variance reduced methods (L-SVRG and SAGA). This optimal minibatch size not only improves the theoretical total complexity of the methods but also improves their convergence in practice, as we show in several experiments.
