Trustworthy deep learning framework for the detection of abnormalities in X-ray shoulder images.

Musculoskeletal conditions affect an estimated 1.7 billion people worldwide, causing intense pain and disability. These conditions lead to 30 million emergency room visits yearly, and the numbers are only increasing. However, diagnosing musculoskeletal issues can be challenging, especially in emergencies where quick decisions are necessary. Deep learning (DL) has shown promise in various medical applications. However, previous methods had poor performance and a lack of transparency in detecting shoulder abnormalities on X-ray images due to a lack of training data and better representation of features. This often resulted in overfitting, poor generalisation, and potential bias in decision-making. To address these issues, a new trustworthy DL framework has been proposed to detect shoulder abnormalities (such as fractures, deformities, and arthritis) using X-ray images. The framework consists of two parts: same-domain transfer learning (TL) to mitigate imageNet mismatch and feature fusion to reduce error rates and improve trust in the final result. Same-domain TL involves training pre-trained models on a large number of labelled X-ray images from various body parts and fine-tuning them on the target dataset of shoulder X-ray images. Feature fusion combines the extracted features with seven DL models to train several ML classifiers. The proposed framework achieved an excellent accuracy rate of 99.2%, F1Score of 99.2%, and Cohen's kappa of 98.5%. Furthermore, the accuracy of the results was validated using three visualisation tools, including gradient-based class activation heat map (Grad CAM), activation visualisation, and locally interpretable model-independent explanations (LIME). The proposed framework outperformed previous DL methods and three orthopaedic surgeons invited to classify the test set, who obtained an average accuracy of 79.1%. The proposed framework has proven effective and robust, improving generalisation and increasing trust in the final results.

Resolving the dynamic properties of entangled linear polymers in non-equilibrium coarse grain simulation with a priori scaling factors.

The molecular weight of polymers can influence the material properties, but the molecular weight at the experiment level sometimes can be a huge burden for property prediction with full-atomic simulations. The traditional bottom-up coarse grain (CG) simulation can reduce the computation cost. However, the dynamic properties predicted by the CG simulation can deviate from the full-atomic simulation result. Usually, in CG simulations, the diffusion is faster and the viscosity and modulus are much lower. The fast dynamics in CG are usually solved by a posteriori scaling on time, temperature, or potential modifications, which usually have poor transferability to other non-fitted physical properties because of a lack of fundamental physics. In this work, a priori scaling factors were calculated by the loss of degrees of freedom and implemented in the iterative Boltzmann inversion. According to the simulation results on 3 different CG levels at different temperatures and loading rates, such a priori scaling factors can help in reproducing some dynamic properties of polycaprolactone in CG simulation more accurately, such as heat capacity, Young's modulus, and viscosity, while maintaining the accuracy in the structural distribution prediction. The transferability of entropy-enthalpy compensation and a dissipative particle dynamics thermostat is also presented for comparison. The proposed method reveals the huge potential for developing customized CG thermostats and offers a simple way to rebuild multiphysics CG models for polymers with good transferability.

Spectroscopic Identification of Charge Transfer of Thiolated Molecules on Gold Nanoparticles via Gold Nanoclusters.

Investigation of charge transfer needs analytical tools that could reveal this phenomenon, and enables understanding of its effect at the molecular level. Here, we show how the combination of using gold nanoclusters (AuNCs) and different spectroscopic techniques could be employed to investigate the charge transfer of thiolated molecules on gold nanoparticles (AuNP@Mol). It was found that the charge transfer effect in the thiolated molecule could be affected by AuNCs, evidenced by the amplification of surface-enhanced Raman scattering (SERS) signal of the molecule and changes in fluorescence lifetime of AuNCs. Density functional theory (DFT) calculations further revealed that AuNCs could amplify the charge transfer process at the molecular level by pumping electrons to the surface of AuNPs. Finite element method (FEM) simulations also showed that the electromagnetic enhancement mechanism along with chemical enhancement determines the SERS improvement in the thiolated molecule. This study provides a mechanistic insight into the investigation of charge transfer at the molecular level between organic and inorganic compounds, which is of great importance in designing new nanocomposite systems. Additionally, this work demonstrates the potential of SERS as a powerful analytical tool that could be used in nanochemistry, material science, energy, and biomedical fields.

Magnetic skyrmions and their manipulations in a 2D multiferroic CuCrP2Te6 monolayer.

Magnetic skyrmions and their effective manipulations are promising for the design of next-generation information storage and processing devices, due to their topologically protected chiral spin textures and low energy cost. They, therefore, have attracted significant interest from the communities of condensed matter physics and materials science. Herein, based on density functional theory (DFT) calculations and micromagnetic simulations, we report the spontaneous 2 nm-diameter magnetic skyrmions in the monolayer CuCrP2Te6 originating from the synergistic effect of broken inversion symmetry and strong Dzyaloshinskii-Moriya interactions (DMIs). The creation and annihilation of magnetic skyrmions can be achieved via the ferroelectric to anti-ferroelectric (FE-to-AFE) transition, due to the variation of the magnetic parameter D2/|KJ|. Moreover, we also found that the DMIs and Heisenberg isotropic exchange can be manipulated by bi-axial strain, to effectively enhance skyrmion stability. Our findings provide feasible approaches to manipulate the skyrmions, which can be used for the design of next-generation information storage devices.

Electrocatalytic Urea Synthesis via N2 Dimerization and Universal Descriptor.

Electrocatalytic urea synthesis through N2 + CO2 coreduction and C-N coupling is a promising and sustainable alternative to harsh industrial processes. Despite considerable efforts, limited progress has been made due to the challenges of breaking inert N≡N bonds for C-N coupling, competing side reactions, and the absence of theoretical principles guiding catalyst design. In this study, we propose a mechanism for highly electrocatalytic urea synthesis using two adsorbed N2 molecules and CO as nitrogen and carbon sources, respectively. This mechanism circumvents the challenging step of N≡N bond breaking and selective CO2 to CO reduction, as the free CO molecule inserts into dimerized *N2 and binds concurrently with two N atoms, forming a specific urea precursor *NNCONN* with both thermodynamic and kinetic feasibility. Through the proposed mechanism, Ti2@C4N3 and V2@C4N3 are identified as highly active catalysts for electrocatalytic urea formation, exhibiting low onset potentials of -0.741 and -0.738 V, respectively. Importantly, taking transition metal atoms anchored on porous graphite-like carbonitride (TM2@C4N3) as prototypes, we introduce a simple descriptor, namely, effective d electron number (Φ), to quantitatively describe the structure-activity relationships for urea formation. This descriptor incorporates inherent atomic properties of the catalyst, such as the number of d electrons, the electronegativity of the metal atoms, and the generalized electronegativity of the substrate atoms, making it potentially applicable to other urea catalysts. Our work advances the comprehension of mechanisms and provides a universal guiding principle for catalyst design in urea electrochemical synthesis.

Ferroelectric Domain and Switching Dynamics in Curved In2Se3: First-Principles and Deep Learning Molecular Dynamics Simulations.

Despite its prevalence in experiments, the influence of complex strain on material properties remains understudied due to the lack of effective simulation methods. Here, the effects of bending, rippling, and bubbling on the ferroelectric domains are investigated in an In2Se3 monolayer by density functional theory and deep learning molecular dynamics simulations. Since the ferroelectric switching barrier can be increased (decreased) by tensile (compressive) strain, automatic polarization reversal occurs in α-In2Se3 with a strain gradient when it is subjected to bending, rippling, or bubbling deformations to create localized ferroelectric domains with varying sizes. The switching dynamics depends on the magnitude of curvature and temperature, following an Arrhenius-style relationship. This study not only provides a promising solution for cross-scale studies using deep learning but also reveals the potential to manipulate local polarization in ferroelectric materials through strain engineering.

Deep Transfer Learning with Enhanced Feature Fusion for Detection of Abnormalities in X-ray Images.

Medical image classification poses significant challenges in real-world scenarios. One major obstacle is the scarcity of labelled training data, which hampers the performance of image-classification algorithms and generalisation. Gathering sufficient labelled data is often difficult and time-consuming in the medical domain, but deep learning (DL) has shown remarkable performance, although it typically requires a large amount of labelled data to achieve optimal results. Transfer learning (TL) has played a pivotal role in reducing the time, cost, and need for a large number of labelled images. This paper presents a novel TL approach that aims to overcome the limitations and disadvantages of TL that are characteristic of an ImageNet dataset, which belongs to a different domain. Our proposed TL approach involves training DL models on numerous medical images that are similar to the target dataset. These models were then fine-tuned using a small set of annotated medical images to leverage the knowledge gained from the pre-training phase. We specifically focused on medical X-ray imaging scenarios that involve the humerus and wrist from the musculoskeletal radiographs (MURA) dataset. Both of these tasks face significant challenges regarding accurate classification. The models trained with the proposed TL were used to extract features and were subsequently fused to train several machine learning (ML) classifiers. We combined these diverse features to represent various relevant characteristics in a comprehensive way. Through extensive evaluation, our proposed TL and feature-fusion approach using ML classifiers achieved remarkable results. For the classification of the humerus, we achieved an accuracy of 87.85%, an F1-score of 87.63%, and a Cohen's Kappa coefficient of 75.69%. For wrist classification, our approach achieved an accuracy of 85.58%, an F1-score of 82.70%, and a Cohen's Kappa coefficient of 70.46%. The results demonstrated that the models trained using our proposed TL approach outperformed those trained with ImageNet TL. We employed visualisation techniques to further validate these findings, including a gradient-based class activation heat map (Grad-CAM) and locally interpretable model-independent explanations (LIME). These visualisation tools provided additional evidence to support the superior accuracy of models trained with our proposed TL approach compared to those trained with ImageNet TL. Furthermore, our proposed TL approach exhibited greater robustness in various experiments compared to ImageNet TL. Importantly, the proposed TL approach and the feature-fusion technique are not limited to specific tasks. They can be applied to various medical image applications, thus extending their utility and potential impact. To demonstrate the concept of reusability, a computed tomography (CT) case was adopted. The results obtained from the proposed method showed improvements.

A Deep Feature Fusion of Improved Suspected Keratoconus Detection with Deep Learning.

Detection of early clinical keratoconus (KCN) is a challenging task, even for expert clinicians. In this study, we propose a deep learning (DL) model to address this challenge. We first used Xception and InceptionResNetV2 DL architectures to extract features from three different corneal maps collected from 1371 eyes examined in an eye clinic in Egypt. We then fused features using Xception and InceptionResNetV2 to detect subclinical forms of KCN more accurately and robustly. We obtained an area under the receiver operating characteristic curves (AUC) of 0.99 and an accuracy range of 97-100% to distinguish normal eyes from eyes with subclinical and established KCN. We further validated the model based on an independent dataset with 213 eyes examined in Iraq and obtained AUCs of 0.91-0.92 and an accuracy range of 88-92%. The proposed model is a step toward improving the detection of clinical and subclinical forms of KCN.

Multifold Fermions and Fermi Arcs Boosted Catalysis in Nanoporous Electride 12CaO·7Al2 O3.

Topological materials have been recently regarded as ideal catalysts for heterogeneous reactions due to their surface metallic states and high carrier mobility. However, the underlying relationship between their catalytic performance and topological states is under debate. It has been discovered that the electride 12CaO·7Al2 O3 (C12A7:4e- ) hosts multifold fermions and Fermi arcs on the (001) surface near the Fermi level due to the interstitial electrons. Through the comparison of catalytic performance under different doping and strain conditions, based on the hydrogen evolution process, it has been demonstrated that the excellent catalytic performance indeed originates from topological properties. A linear relationship between the length of Fermi arcs, and Gibbs free energy (ΔGH* ) has been found, which not only provides the direct evidence to link the enhanced catalytic performance and surface Fermi arc states, but also fully clarifies the fundamental mechanism in topological catalysis.

Feasibility of uranium tailings for cemented backfill and its environmental effects.

Surface storage of uranium tailings presents a potential threat to the environment and human health. Cemented backfill can be used to dispose of tailings and control the ground pressure of stopes, providing a new approach for the in-situ seal of heap leaching uranium tailings (HLUTs). The backfilling characteristics of HLUTs were investigated by analyzing the release mechanism of sulfuric acid in HLUTs, the rheological properties of backfill slurry, as well as the strength development and microscopic characteristics of cemented HLUTs backfill (CUTB). The environmental effects of the CUTB were also assessed, and a novel filling process was presented. The results showed that the release rate of sulfuric acid in HLUTs decreased logarithmically, and the content of free sulfuric acid in coarse particles surfaces and ultrafine particles is high, which can be pretreated with 0.1 % quicklime. Slurry with a mass concentration of 74 % ~ 76 % can satisfy the requirements for pipeline transport. The CUTB's strength raised quickly in the former 90d, then decreased to a different extent after 150d, adding 50 wt% FA can enhance its later stability. The leaching level of uranium in CUTB cured for 28d is below the stipulated limit (GB 23727-2009) under different test conditions, having little impact on the underground environment. The hydration products of CUTB are mainly gypsum and C-S-H gel. Gypsum causes later degradation in strength; numerous C-S-H gels generated by the secondary hydration of FA enhance the resistance to sulfate corrosion. These findings have demonstrated that cemented backfill has a high inclusion ratio and low cost for HLUTs, which is of great significance to the HLUTs minimization and the safety of mining while promoting the environmentally friendly development of uranium mines and mills.

Volcano relationships and a new activity descriptor of 2D transition metal-Fe layered double hydroxides for efficient oxygen evolution reaction.

Iron (Fe) sites play a critical role in boosting the catalytic activity of transition metal layered double hydroxide (LDH) electrocatalysts for the oxygen evolution reaction (OER), but the contribution of the Fe content to the catalysis of Fe-doped LDHs is still not well understood. Herein, a series of two-dimensional (2D) Fe-doped MFe-LDHs (M = Co, Ni, Cu, and Mn) was synthesized via a general molecular self-assembly method to track the role of Fe in their electrocatalytic OER activities. Besides the revelation of the intrinsic activity trend of NiFe > CoFe > MnFe > CuFe, volcano-shaped relationships among the catalytic activity descriptors, i.e., overpotential, Tafel slope, and turnover frequency (TOF), and the Fe-content in MFe-LDHs, were identified. Specifically, a ∼20% Fe content resulted in the highest OER performance for the LDH, while excess Fe compromised its activity. A similar volcano relationship was determined between the intermediate adsorption and Fe content via operando impedance spectroscopy (EIS) measurements, and it was shown that the intermediate adsorption capacitance (CPEad) can be a new activity descriptor for electrocatalysts. In this work, we not only performed a systematic study on the role of Fe in 2D Fe-doped LDHs but also offer some new insights into the activity descriptors for electrocatalysts.

Effect of nonaffine displacement on the mechanical performance of degraded PCL and its graphene composites: an atomistic investigation.

Evaluating the mechanical properties of biodegradable implants can be challenging for in situ experiments and time-consuming for materials with a slow degradation rate, such as polycaprolactone (PCL). In this work, the effects of chain scission and water erosion on the mechanical properties of degraded PCL are investigated by molecular dynamics simulation. The decrease of the mechanical performance is correlated with the increase of the nonaffine displacement during the degradation. The nonaffine squared displacements (NSD) during the tensile deformation are calculated by subtracting the affine squared displacements from the mean squared displacements. After chain scission, short polymer chains increase the NSD of the system and weaken the modulus of the polymer matrix. The effect of the NSD is also observed in a water erosion model. When the bond break ratio is less than 5%, PCL still maintains a well-entangled network, which constrains the diffusion of the water molecules, resulting in a higher modulus of the erosion model than the chain scission model at a low degradation rate. The effect of NSD is also found in the PCL/graphene composites. For the degraded polymer chains, the diffusion of PCL is constrained by the graphene network, and such an effect increases during the degradation. As a result, the addition of graphene nanosheets slows down the decreasing trend of Young's modulus. Such findings can also explain the size effect of the graphene reinforcement on the mechanical properties of the polymer composites. This work provides atomistic insights into the mechanical property evolution during polymer degradation, revealing the possibility of tuning the mechanical performance by controlling the diffusion, which could be beneficial for the design and lifetime prediction of degradable implants.

Nanoscale Diamane Spiral Spring for High Mechanical Energy Storage.

A compact, stable, sustainable, and high-energy density power supply system is crucial for the engineering deployment of mobile electromechanical devices/systems either at the small- or large-scale. This work proposes a spiral-based mechanical energy storage scheme utilizing the newly synthesized 2D diamane. Atomistic simulations show that diamane spiral can achieve a high theoretical gravimetric energy density of about 564 Wh kg-1 , about 14 500 times the steel spring. The interlayer friction between diamane is found to cause a strong stick-slip effect that results in local stress/strain concentration. As such, the energy storage capacity of the diamane spiral can be tuned by suppressing the influence from the interlayer friction. Simulations affirm that higher gravimetric energy density can be achieved by reducing the turn number or adopting a low friction contact pair. The fundamental principles that dominate the energy storage capacity of the spiral spring are theoretically analyzed, respectively. The obtained insights suggest that the 2D vdW solids can be promising candidates to construct spiral structures with a high gravimetric energy density. This work should be beneficial for the design of reliable, stable, and sustainable nanoscale mechanical energy storage schemes that can be used as an alternative low-carbon footage energy supplier for novel micro-/nanoscale devices or systems.

Design optimization of a magnesium-based metal hydride hydrogen energy storage system.

Metal hydrides (MH) are known as one of the most suitable material groups for hydrogen energy storage because of their large hydrogen storage capacity, low operating pressure, and high safety. However, their slow hydrogen absorption kinetics significantly decreases storage performance. Faster heat removal from MH storage can play an essential role to enhance its hydrogen absorption rate, resulting in better storage performance. In this regard, the present study aims to improve heat transfer performance to positively impact the hydrogen absorption rate of MH storage systems. A novel semi-cylindrical coil is first designed and optimized for hydrogen storage and embedded as an internal heat exchanger with air as the heat transfer fluid (HTF). The effect of novel heat exchanger configurations is analyzed and compared with normal helical coil geometry, based on various pitch sizes. Furthermore, the operating parameters of MH storage and HTF are numerically investigated to obtain optimal values. ANSYS Fluent 2020 R2 is utilized for the numerical simulations. Results from this study demonstrate that MH storage performance is significantly improved by using a semi-cylindrical coil heat exchanger (SCHE). The hydrogen absorption duration reduces by 59% compared to a normal helical coil heat exchanger. The lowest coil pitch from SCHE leads to a 61% reduction of the absorption time. In terms of operating parameters for the MH storage with SCHE, all selected parameters provide a major improvement in the hydrogen absorption process, especially the inlet temperature of the HTF.

Heat Wave and Bushfire Meteorology in New South Wales, Australia: Air Quality and Health Impacts.

The depletion of air quality is a major problem that is faced around the globe. In Australia, the pollutants emitted by bushfires play an important role in making the air polluted. These pollutants in the air result in many adverse impacts on the environment. This paper analysed the air pollution from the bushfires from November 2019 to July 2020 and identified how it affects the human respiratory system. The bush fires burnt over 13 million hectares, destroying over 2400 buildings. While these immediate effects were devastating, the long-term effects were just as devastating, with air pollution causing thousands of people to be admitted to hospitals and emergency departments because of respiratory complications. The pollutant that caused most of the health effects throughout Australia was Particulate Matter (PM) PM2.5 and PM10. Data collection and analysis were covered in this paper to illustrate where and when PM2.5 and PM10, and other pollutants were at their most concerning levels. Susceptible areas were identified by analysing environmental factors such as temperature and wind speed. The study identified how these pollutants in the air vary from region to region in the same time interval. This study also focused on how these pollutant distributions vary according to the temperature, which helps to determine the relationship between the heatwave and air quality. A computational model for PM2.5 aerosol transport to the realistic airways was also developed to understand the bushfire exhaust aerosol transport and deposition in airways. This study would improve the knowledge of the heat wave and bushfire meteorology and corresponding respiratory health impacts.

Molecular basis of transport of surface functionalised gold nanoparticles to pulmonary surfactant.

Ligands like alkanethiol (e.g. dodecanethiol, hexadecanethiol, etc.) and polymers (e.g. poly(vinyl pyrrolidone), polyethylene glycol-thiol) capped to the gold nanoparticles (AuNPs) are widely used in biomedical field as drug carriers and as promising materials for probing and manipulating cellular processes. Ligand functionalised AuNPs are known to interact with the pulmonary surfactant (PS) monolayer once reaching the alveolar region. Therefore, it is crucial to understand the interaction between AuNPs and PS monolayers. Using coarse-grained molecular dynamics simulations, the effect of ligand density, and ligand length have been studied for two classes of ligands on a PS model monolayer consisting of DPPC, POPG, cholesterol and SP-B (mini-peptide). The ligands considered in this study are alkanethiol and polyethylene glycol (PEG) thiol as examples of hydrophobic and hydrophilic ligands, respectively. It was observed that the interaction between AuNPs and PS changes the biophysical properties of PS monolayer in compressed and expanded states. The AuNPs with hydrophilic ligand, can penetrate through the monolayer more easily, while the AuNPs with hydrophobic ligand are embedded in the monolayer and participated in deforming the monolayer structure particularly the monolayer in the compressed state. The bare AuNPs hinder to lower the monolayer surface tension value at the interface, however introducing ligand to the bare AuNPs or increasing the ligand length and density have an impact of lowering of monolayer surface tension to a minor extent. The simulation results guide the design of ligand protected NPs as drug carriers and can identify the nanoparticles' potential side effects on lung surfactant.

Is the SARS CoV-2 Omicron Variant Deadlier and More Transmissible Than Delta Variant?

Genetic variants of severe acute respiratory syndrome coronavirus (SARS-CoV-2) have been globally surging and devastating many countries around the world. There are at least eleven reported variants dedicated with inevitably catastrophic consequences. In 2021, the most dominant Delta and Omicron variants were estimated to lead to more severity and deaths than other variants. Furthermore, these variants have some contagious characteristics involving high transmissibility, more severe illness, and an increased mortality rate. All outbreaks caused by the Delta variant have been rapidly skyrocketing in infection cases in communities despite tough restrictions in 2021. Apart from it, the United States, the United Kingdom and other high-rate vaccination rollout countries are still wrestling with this trend because the Delta variant can result in a significant number of breakthrough infections. However, the pandemic has changed since the latest SARS-CoV-2 variant in late 2021 in South Africa, Omicron. The preliminary data suggest that the Omicron variant possesses 100-fold greater than the Delta variant in transmissibility. Therefore, this paper aims to review these characteristics based on the available meta-data and information from the first emergence to recent days. Australia and the five most affected countries, including the United States, India, Brazil, France, as well as the United Kingdom, are selected in order to review the transmissibility, severity and fatality due to Delta and Omicron variants. Finally, the vaccination programs for each country are also reviewed as the main factor in prevention.

Conversion of Catalytically Inert 2D Bismuth Oxide Nanosheets for Effective Electrochemical Hydrogen Evolution Reaction Catalysis via Oxygen Vacancy Concentration Modulation.

Oxygen vacancies (Vo) in electrocatalysts are closely correlated with the hydrogen evolution reaction (HER) activity. The role of vacancy defects and the effect of their concentration, however, yet remains unclear. Herein, Bi2O3, an unfavorable electrocatalyst for the HER due to a less than ideal hydrogen adsorption Gibbs free energy (ΔGH*), is utilized as a perfect model to explore the function of Vo on HER performance. Through a facile plasma irradiation strategy, Bi2O3 nanosheets with different Vo concentrations are fabricated to evaluate the influence of defects on the HER process. Unexpectedly, while the generated oxygen vacancies contribute to the enhanced HER performance, higher Vo concentrations beyond a saturation value result in a significant drop in HER activity. By tunning the Vo concentration in the Bi2O3 nanosheets via adjusting the treatment time, the Bi2O3 catalyst with an optimized oxygen vacancy concentration and detectable charge carrier concentration of 1.52 × 1024 cm-3 demonstrates enhanced HER performance with an overpotential of 174.2 mV to reach 10 mA cm-2, a Tafel slope of 80 mV dec-1, and an exchange current density of 316 mA cm-2 in an alkaline solution, which approaches the top-tier activity among Bi-based HER electrocatalysts. Density-functional theory calculations confirm the preferred adsorption of H* onto Bi2O3 as a function of oxygen chemical potential (∆µO) and oxygen partial potential (PO2) and reveal that high Vo concentrations result in excessive stability of adsorbed hydrogen and hence the inferior HER activity. This study reveals the oxygen vacancy concentration-HER catalytic activity relationship and provides insights into activating catalytically inert materials into highly efficient electrocatalysts.

Machine learning-based modeling in food processing applications: State of the art.

Food processing is a complex, multifaceted problem that requires substantial human interaction to optimize the various process parameters to minimize energy consumption and ensure better-quality products. The development of a machine learning (ML)-based approach to food processing applications is an exciting and innovative idea for optimizing process parameters and process kinetics to reduce energy consumption, processing time, and ensure better-quality products; however, developing such a novel approach requires significant scientific effort. This paper presents and evaluates ML-based approaches to various food processing operations such as drying, frying, baking, canning, extrusion, encapsulation, and fermentation to predict process kinetics. A step-by-step procedure to develop an ML-based model and its practical implementation is presented. The key challenges of neural network training and testing algorithms and their limitations are discussed to assist readers in selecting algorithms for solving problems specific to food processing. In addition, this paper presents the potential and challenges of applying ML-based techniques to hybrid food processing operations. The potential of physics-informed ML modeling techniques for food processing applications and their strategies is also discussed. It is expected that the potential information of this paper will be valuable in advancing the ML-based technology for food processing applications.

Stacking-Dependent Interlayer Ferroelectric Coupling and Moiré Domains in a Twisted AgBiP2Se6 Bilayer.

Rotation/twisting of bilayers could induce unprecedented new physics due to stacking-dependent electronic properties and interlayer coupling, such as the superconductivity in twisted bilayer graphene, which can find applications in electronics. However, deep understanding at the atomic/electronic levels is limited by the capability of accurate theoretical simulations. Here, from first-principles simulations, we found that the AgBiP2Se6 bilayer has stacking-dependent ferroelectric ground states due to interlayer polarization coupling. Interlayer ferroelectric coupling is preferred in an AA-stacked AgBiP2Se6 bilayer, but antiferroelectric coupling is preferred in AB- or AC-stacked configurations. The ferroelectric Moiré patterns are thus observed in a twisted AgBiP2Se6 bilayer with ferroelectric (antiferroelectric) interlayer couplings in the AA (AB/AC)-stacked areas. Our work for the first time unveils the effects of twisting/rotation on interlayer polarization coupling and provides a real example of ferroelectric Moiré patterns.