Search

This paper introduces a design of ternary bistale ring physical unclonable function (BRPUF) based on carbon nanotube field effect transistors (CNTFETs). PUFs can extract unique chip features from process variations and assign every chip an inherent fingerprint to be used for identification or authentication of devices. As the main point of strength for PUF security, the size of challenge-response pairs (CRPs) can be increased by adopting multi-valued logic designs. The proposed ternary BRPUF blends higher entropy into CRPs and is tested to produce unbiased response digits, thus making them more unpredictable and more resistant against machine learning attacks. The instinctive features of CNTFETs, such as intensive process variations and high tolerance against temperature fluctuations and noises, enable our design to function robustly in a wide range of operating conditions. The proposed design has been implemented using standard 32-nm CNTFET model libraries and tested by HSPICE simulations. The results elaborate the superiority of our work compared with the traditional binary BRPUFs. [ABSTRACT FROM AUTHOR]

Ternary polymer solar cells (PSCs) are currently the simplest and most efficient way to further improve the device performance in PSCs. To find high-performance organic photovoltaic materials, the established connection between the material structure and device performance before fabrication is of great significance. Herein, firstly, a database of the photovoltaic performance in 874 experimental PSCs reported in the literature is established, and three different fingerprint expressions of a molecular structure are explored as input features; the results show that long fingerprints of 2D atom pairs can contain more effective information and improve the accuracy of the models. Through supervised learning, five machine learning (ML) models were trained to build a mapping of the photovoltaic performance improvement relationship from binary to ternary PSCs. The GBDT model had the best predictive ability and generalization. Eighteen key structural features from a non-fullerene acceptor and the third components that affect the device's PCE were screened based on this model, including a nitrile group with lone-pair electron, a halogen atom, an oxygen atom, etc. Interestingly, the structural features for the enhanced device's PCE were essentially increased by the Jsc or FF. More importantly, the reliability of the ML model was further verified by preparing the highly efficient PSCs. Taking the PM6:BTP-eC9:PY-IT ternary PSC as an example, the PCE prediction (18.03%) by the model was in good agreement with the experimental results (17.78%), the relative prediction error was 1.41%, and the relative error between all experimental results and predicted results was less than 5%. These results indicate that ML is a useful tool for exploring the photovoltaic performance improvement of PSCs and accelerating the design and application with highly efficient non-fullerene materials. [ABSTRACT FROM AUTHOR]

Electromagnetic methods for non-destructive evaluation (NDE) are presented, with which sheet metal components can be identified and their material properties can be characterized. The latter is possible with 3MA, the Micromagnetic Multiparametric Microstructure and stress Analyser. This is a combination of several micromagnetic NDE methods that make it possible to analyse the microstructure in a ferromagnetic material and to determine quantitative values of the mechanical material properties or the stress state. In the case of cold forming, the 3MA application for pre-process testing of sheet metal is discussed. Based on the 3MA information, the formability of the sheets can be predicted. To apply 3MA in-line, the influence of the relative speed and the relative distance between the 3MA probe head and the sheet was investigated. In a second study, a spatially resolved eddy current (EC) method was used to create an image of the intrinsic material microstructure of a component for its identification and traceability. It turned out, that these intrinsic fingerprint images can still be recognized even after subsequent plastic deformation or coating of the surface. This enabled the development of a marker-free traceability method for sheet metal processing. It is based on a low-cost array sensor and a specimen identification using robust and partly redundant features of the fingerprint images processed by machine learning (ML). [ABSTRACT FROM AUTHOR]

Background The flavor theory of Chinese herbal medicines (CHMs) is one of the core theories of traditional Chinese medicine (TCM). Accurate flavor identification of CHMs is essential to guide the clinical application of CHMs. Objective To develop a new method for flavor identification of CHMs according to the ingredient information for CHMs. Methods It was found that the chemical basis of medicinal flavors was CHM ingredients. We developed a bitter–pungent flavor identification scheme to build a relationship between medicinal flavors and CHM ingredients. We firstly proposed a scientific hypothesis that "CHMs with similar flavors should have a similar chemical basis". To test this scientific hypothesis, we then explored an intelligent algorithm for bitter-pungent flavor identification of CHMs based on the information similarity of CHM ingredients. GC was used to separate the chemical ingredients of CHMs and analyze the ingredient information of CHMs. A distance metric learning algorithm was built to measure the similarity of GC chemical fingerprints. A bitter-pungent flavor identification scheme (BPFI) was proposed to predict the bitter–pungent flavor of CHMs. Finally, a number of experiments were performed to evaluate the identification performance of our scheme. Results Compared to classical algorithms, our proposed BPFI scheme has better flavor prediction performance. The total identification accuracy of our BPFI scheme reached 0.843. The area under ROC (receiver operating characteristic curve) curve (AUC) was 0.899. Conclusion The experimental results confirmed our inference that the chemical basis of CHM flavors was CHM ingredients, and implied that CHMs with similar flavors had similar composition. The BPFI model proved to be effective and feasible. Highlights Verification hypothesis: CHMs with similar flavors should have similar chemical basis. [ABSTRACT FROM AUTHOR]

Monoamine oxidases (MAOs), specifically MAO-A and MAO-B, play important roles in the breakdown of monoamine neurotransmitters. Therefore, MAO inhibitors are crucial for treating various neurodegenerative disorders, including Parkinson's disease (PD), Alzheimer's disease (AD), and amyotrophic lateral sclerosis (ALS). In this study, we developed a novel cheminformatics pipeline by generating three diverse molecular feature-based machine learning-assisted quantitative structural activity relationship (ML-QSAR) models concerning MAO-B inhibition. PubChem fingerprints, substructure fingerprints, and one-dimensional (1D) and two-dimensional (2D) molecular descriptors were implemented to unravel the structural insights responsible for decoding the origin of MAO-B inhibition in 249 non-reductant molecules. Based on a random forest ML algorithm, the final PubChem fingerprint, substructure fingerprint, and 1D and 2D molecular descriptor prediction models demonstrated significant robustness, with correlation coefficients of 0.9863, 0.9796, and 0.9852, respectively. The significant features of each predictive model responsible for MAO-B inhibition were extracted using a comprehensive variance importance plot (VIP) and correlation matrix analysis. The final predictive models were further developed as a web application, MAO-B-pred (https://mao-b-pred.streamlit.app/), to allow users to predict the bioactivity of molecules against MAO-B. Molecular docking and dynamics studies were conducted to gain insight into the atomic-level molecular interactions between the ligand-receptor complexes. These findings were compared with the structural features obtained from the ML-QSAR models, which supported the mechanistic understanding of the binding phenomena. The presented models have the potential to serve as tools for identifying crucial molecular characteristics for the rational design of MAO-B target inhibitors, which may be used to develop effective drugs for neurodegenerative disorders. [ABSTRACT FROM AUTHOR]

In recent years, considerable and valuable research progress has been made in indoor positioning technologies based on WLAN Radio Frequency (RF) fingerprinting, identifying it as one of the most promising positioning technologies with substantial potential for wider adoption. However, indoor environmental factors significantly influence the propagation of wireless RF signals, resulting in a considerable decrease in positioning accuracy as the indoor environmental conditions vary. Thus, effectively mitigating the impact of indoor environmental factors on WLAN RF fingerprinting-based positioning systems has become a crucial research problem. Currently, there is a dearth of comprehensive research on the influence of indoor climatic factors, particularly the variations in relative humidity, on the propagation of WLAN RF signals within indoor spaces and its consequential impact on positioning accuracy. To address the aforementioned issues, this paper proposes an Adaptive expansion fingerprint database (AeFd) model based on a regression learning algorithm. The AeFd, through the design of a relationship model describing the interaction between fingerprint databases under varying relative humidity, allows the fingerprint database expanded by AeFd to dynamically adapt to the changes in indoor relative humidity. Our experiments show that using the AeFd model with the KNN algorithm, a 5% performance improvement was observed over 10 days and an 8% improvement over 10 months. According to experimental test results, the fingerprint database expansion model AeFd proposed in this paper can effectively expand the fingerprint database under different relative humidity levels, thereby significantly enhancing the positioning performance of the system and improving its stability. [ABSTRACT FROM AUTHOR]

The article provides a comprehensive survey on the use of Graph Neural Networks (GNNs) in recommender systems (RSs), highlighting their role in enhancing recommendations through advanced learning techniques. Topics discussed include GNN-based recommender models, challenges in data quality and computational efficiency, and future research directions such as explainability, scalability, and dynamic graph handling.

The integration of drug molecular representations into predictive models for Drug Response Prediction (DRP) is a standard procedure in pharmaceutical research and development. However, the comparative effectiveness of combining these representations with genetic profiles for DRP remains unclear. This study conducts a comprehensive evaluation of the efficacy of various drug molecular representations employing cutting-edge machine learning models under various experimental settings. Our findings reveal that the inclusion of molecular representations from either PubChem fingerprints or SMILES can significantly enhance the performance of DRPs when used in conjunction with deep learning models. However, the optimal choice of drug molecular representation can vary depending on the predictive model and the specific DRP task. The insights derived from our study offer useful guidance on selecting the most suitable drug molecular representations for constructing efficient predictive models for DRPs, aiding for drug repurposing, personalized medicine, and new drug discovery. [ABSTRACT FROM AUTHOR]

Fingerprints hold paramount importance in various fields due to their unique characteristics, making them invaluable for identification and verification purposes. This paper presents a state-of-the-art methodology for fingerprint recognition leveraging the Hybrid Deep Learning Model - SpatioTemporalNet (STNet), enhanced by Hybrid GAPSO Optimization for Feature Matching. By fusing Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and advanced optimization techniques, this approach revolutionizes the landscape of fingerprint analysis. STNet integrates both static and dynamic feature analysis, enabling adaptability to diverse conditions encountered in real-world scenarios. The integration of STNet and Hybrid GAPSO Optimization represents a significant advancement in fingerprint recognition technology, offering unparalleled levels of security and reliability. By harnessing the power of deep learning and optimization, the proposed methodology excels in handling various challenges such as partial prints, distortions, and variations in fingerprint impressions. Implemented using Python, the proposed methodology undergoes rigorous experimentation, demonstrating exceptional performance with an accuracy rate of 98.7%. This high level of accuracy underscores the effectiveness and reliability of the approach in accurately identifying individuals based on their unique fingerprint patterns. [ABSTRACT FROM AUTHOR]

Extended-connectivity fingerprints (ECFPs) are a ubiquitous tool in current cheminformatics and molecular machine learning, and one of the most prevalent molecular feature extraction techniques used for chemical prediction. Atom features learned by graph neural networks can be aggregated to compound-level representations using a large spectrum of graph pooling methods. In contrast, sets of detected ECFP substructures are by default transformed into bit vectors using only a simple hash-based folding procedure. We introduce a general mathematical framework for the vectorisation of structural fingerprints via a formal operation called substructure pooling that encompasses hash-based folding, algorithmic substructure selection, and a wide variety of other potential techniques. We go on to describe Sort & Slice, an easy-to-implement and bit-collision-free alternative to hash-based folding for the pooling of ECFP substructures. Sort & Slice first sorts ECFP substructures according to their relative prevalence in a given set of training compounds and then slices away all but the L most frequent substructures which are subsequently used to generate a binary fingerprint of desired length, L. We computationally compare the performance of hash-based folding, Sort & Slice, and two advanced supervised substructure-selection schemes (filtering and mutual-information maximisation) for ECFP-based molecular property prediction. Our results indicate that, despite its technical simplicity, Sort & Slice robustly (and at times substantially) outperforms traditional hash-based folding as well as the other investigated substructure-pooling methods across distinct prediction tasks, data splitting techniques, machine-learning models and ECFP hyperparameters. We thus recommend that Sort & Slice canonically replace hash-based folding as the default substructure-pooling technique to vectorise ECFPs for supervised molecular machine learning. Scientific contribution A general mathematical framework for the vectorisation of structural fingerprints called substructure pooling; and the technical description and computational evaluation of Sort & Slice, a conceptually simple and bit-collision-free method for the pooling of ECFP substructures that robustly and markedly outperforms classical hash-based folding at molecular property prediction. [ABSTRACT FROM AUTHOR]

This study illustrates the use of chemical fingerprints with machine learning for blood–brain barrier (BBB) permeability prediction. Employing the Blood Brain Barrier Database (B3DB) dataset for BBB permeability prediction, we extracted nine different fingerprints. Support Vector Machine (SVM) and Extreme Gradient Boosting (XGBoost) algorithms were used to develop models for permeability prediction. Random Forest recursive Feature Selection (RF-RFS) method was used for extracting informative attributes. An additional database was employed for the validation phase. The results indicate that all nine datasets achieved good performance in training, test and validation stages. We further took MACC Keys fingerprints, one of the best performing models for explainability analysis. For this purpose, we used SHapley Additive exPlanations (SHAP) analysis on this dataset for the identification of key structural features influencing BBB permeability prediction. These features include aliphatic carbons, methyl groups and oxygen-containing groups. This study highlights the effectiveness of different fingerprint descriptors in predicting BBB permeability. SHAP analysis provides value additions to the simulations. These simulations will be of significant help in drug discovery processes, particularly in developing Central Nervous System (CNS) therapeutics. [ABSTRACT FROM AUTHOR]

In quest of improving the productivity and efficiency of manufacturing processes, Artificial Intelligence (AI) is being used extensively for response prediction, model dimensionality reduction, process optimization, and monitoring. Though having superior accuracy, AI predictions are unintelligible to the end users and stakeholders due to their opaqueness. Thus, building interpretable and inclusive machine learning (ML) models is a vital part of the smart manufacturing paradigm to establish traceability and repeatability. The study addresses this fundamental limitation of AI-driven manufacturing processes by introducing a novel Explainable AI (XAI) approach to develop interpretable processes and product fingerprints. Here the explainability is implemented in two stages: by developing interpretable representations for the fingerprints, and by posthoc explanations. Also, for the first time, the concept of process fingerprints is extended to develop an interpretable probabilistic model for bottleneck events during manufacturing processes. The approach is demonstrated using two datasets: nanosecond pulsed laser ablation to produce superhydrophobic surfaces and wire EDM real-time monitoring dataset during the machining of Inconel 718. The fingerprint identification is performed using a global Lipschitz functions optimization tool (MaxLIPO) and a stacked ensemble model is used for response prediction. The proposed interpretable fingerprint approach is robust to change in processes and can responsively handle both continuous and categorical responses alike. Implementation of XAI not only provided useful insights into the process physics but also revealed the decision-making logic for local predictions. [ABSTRACT FROM AUTHOR]

In this study, it is aimed to establish a novel method based on a deep‐learning‐guided surface‐enhanced Raman spectroscopy (SERS) technique to achieve rapid and accurate classification of vaginal cleanliness levels. We proposed a variational autoencoder (VAE) approach to enhance spectral quality, coupled with a deep learning algorithm long short‐term memory (LSTM) neural network to analyze SERS spectra produced by vaginal secretions. The performance of various machine learning (ML) algorithms is assessed using multiple evaluation metrics. Finally, the reliability of the optimal model is tested using blind test data (N = 10/group for each cleanliness level). The data quality of the SERS fingerprints of four types of vaginal secretions is significantly improved after VAE decoding and reconstruction. The signal‐to‐noise ratio of the generated spectra increased from the original 2.58–11.13. Among all algorithms, the VAE–LSTM algorithm demonstrates the best prediction ability and time efficiency. Additionally, blind test datasets yielded an overall accuracy of 85%. In this study, it is concluded that the deep‐learning‐guided SERS technique holds significant potential in rapidly distinguishing between different levels of vaginal cleanliness through human vaginal secretion samples. This contributes to the efficient diagnosis of vaginal cleanliness levels in clinical settings. [ABSTRACT FROM AUTHOR]

Currently, a significant number of pre-trained models are published online to provide services to users owing to the rapid maturation and popularization of machine learning as a service (MLaaS). Some malicious users have pre-trained models illegally to redeploy them and earn money. However, most of the current methods focus on verifying the copyright of the model rather than tracing responsibility for the suspect model. In this study, TraceGuard is proposed, the first framework based on steganography for tracing a suspect self-supervised learning (SSL) pre-trained model, to ascertain which authorized user illegally released the suspect model or if the suspect model is independent. Concretely, the framework contains an encoder and decoder pair and the SSL pre-trained model. Initially, the base pre-trained model is frozen, and the encoder and decoder are jointly learned to ensure the two modules can embed the secret key into the cover image and extract the secret key from the embedding output by the base pre-trained model. Subsequently, the base pre-trained model is fine-tuned using stego images to implement a fingerprint while the encoder and decoder are frozen. To assure the effectiveness and robustness of the fingerprint and the utility of fingerprinted pre-trained models, three alternate steps of model stealing simulations, fine-tuning for uniqueness, and fine-tuning for utility are designed. Finally, the suspect pre-trained model is traced to its user by querying stego images. Experimental results demonstrate that TraceGuard can reliably trace suspect models and is robust against common fingerprint removal attacks such as fine-tuning, pruning, and model stealing. In the future, we will further improve the robustness against model stealing attack. [ABSTRACT FROM AUTHOR]

Machine learning (ML) is effective to establish the complicated trilateral relationship among structures, properties, and photovoltaic performance, which is fundamental issue in developing novel materials for improving power conversion efficiency (PCE) of organic solar cells (OSCs). Herein, we constructed the database of 397 donor–acceptor pairs of OSCs with photovoltaic parameters and descriptor sets, which include donor–acceptor weight ratio within the active layer of the OSCs, root mean square of roughness, and 1024-bit Morgan molecular fingerprint for donor (Fp-D) and acceptor (Fp-A). The ML models random forest (RF), adaptive boosting (AdaBoost), extra trees regression, and gradient boosting regression trees were trained based on the descriptor set. The metrics determination coefficient (R2), Pearson correlation coefficient (r), root mean square error, and mean absolute error were selected to evaluate ML model performances. The results showed that the RF model exhibits the highest accuracy and stability for PCE prediction among these four ML models. Moreover, based on the decomposition of non-fullerene acceptors L8-BO, BTP-ec9, AQx-2, and IEICO, 20 acceptor molecules with symmetric A–D–A and A–π–D–π–A architectures were designed. The photovoltaic parameters of the designed acceptors were predicted using the trained RF model, and the virtual screening of designed acceptors was conducted based on the predicted PCE. The results indicate that six designed acceptors can reach the predicted PCE higher than 12% when P3HT was adopted as a donor. While PM6 was applied as a donor, five designed acceptors can achieve the predicted PCE higher than 16%. [ABSTRACT FROM AUTHOR]

Conventional machine learning (ML) and deep learning (DL) play a key role in the selectivity prediction of kinase inhibitors. A number of models based on available datasets can be used to predict the kinase profile of compounds, but there is still controversy about the advantages and disadvantages of ML and DL for such tasks. In this study, we constructed a comprehensive benchmark dataset of kinase inhibitors, involving in 141,086 unique compounds and 216,823 well-defined bioassay data points for 354 kinases. We then systematically compared the performance of 12 ML and DL methods on the kinase profiling prediction task. Extensive experimental results reveal that (1) Descriptor-based ML models generally slightly outperform fingerprint-based ML models in terms of predictive performance. RF as an ensemble learning approach displays the overall best predictive performance. (2) Single-task graph-based DL models are generally inferior to conventional descriptor- and fingerprint-based ML models, however, the corresponding multi-task models generally improves the average accuracy of kinase profile prediction. For example, the multi-task FP-GNN model outperforms the conventional descriptor- and fingerprint-based ML models with an average AUC of 0.807. (3) Fusion models based on voting and stacking methods can further improve the performance of the kinase profiling prediction task, specifically, RF::AtomPairs + FP2 + RDKitDes fusion model performs best with the highest average AUC value of 0.825 on the test sets. These findings provide useful information for guiding choices of the ML and DL methods for the kinase profiling prediction tasks. Finally, an online platform called KIPP (https://kipp.idruglab.cn) and python software are developed based on the best models to support the kinase profiling prediction, as well as various kinase inhibitor identification tasks including virtual screening, compound repositioning and target fishing. [ABSTRACT FROM AUTHOR]

This research work focuses on a Near-Infra-Red (NIR) finger-images-based multimodal biometric system based on Finger Texture and Finger Vein biometrics. The individual results of the biometric characteristics are fused using a fuzzy system, and the final identification result is achieved. Experiments are performed for three different databases, i.e., the Near-Infra-Red Hand Images (NIRHI), Hong Kong Polytechnic University (HKPU) and University of Twente Finger Vein Pattern (UTFVP) databases. First, the Finger Texture biometric employs an efficient texture feature extracting algorithm, i.e., Linear Binary Pattern. Then, the classification is performed using Support Vector Machine, a proven machine learning classification algorithm. Second, the transfer learning of pre-trained convolutional neural networks (CNNs) is performed for the Finger Vein biometric, employing two approaches. The three selected CNNs are AlexNet, VGG16 and VGG19. In Approach 1, before feeding the images for the training of the CNN, the necessary preprocessing of NIR images is performed. In Approach 2, before the pre-processing step, image intensity optimization is also employed to regularize the image intensity. NIRHI outperforms HKPU and UTFVP for both of the modalities of focus, in a unimodal setup as well as in a multimodal one. The proposed multimodal biometric system demonstrates a better overall identification accuracy of 99.62% in comparison with 99.51% and 99.50% reported in the recent state-of-the-art systems. [ABSTRACT FROM AUTHOR]

Modern solutions for precise fault localisation in Low Voltage (LV) Distribution Networks (DNs) often rely on costly tools such as the micro-Phasor Measurement Unit (μ PMU), which is potentially impractical for the large number of nodes in LVDNs. This paper introduces a novel fault detection technique using a distribution network digital twin without the use of μ PMUs. The Digital Twin (DT) integrates data from Smart Meters (SMs) and network topology to create an accurate replica. In using SM voltage-magnitude readings, the pre-built twin compiles a database of fault scenarios and matches them with their unique voltage fingerprints. However, this SM-based voltage-only approach shows only a 70.7% accuracy in classifying fault type and location. Therefore, this research suggests using the cables' Currents Symmetrical Component (CSC). Since SMs do not provide direct current data, a Machine Learning (ML)-based regression method is proposed to estimate the cables' currents in the DT. Validation is performed on a 41-node LV distribution feeder in the Scottish network provided by the industry partner Scottish Power Energy Networks (SPEN). The results show that the current estimation regressor significantly improves fault localisation and identification accuracy to 95.77%. This validates the crucial role of a DT in distribution networks, thus enabling highly accurate fault detection when using SM voltage-only data, with further refinements being conducted through estimations of CSC. The proposed DT offers automated fault detection, thus enhancing customer connectivity and maintenance team dispatch efficiency without the need for additional expensive μ PMU on a densely-noded distribution network. [ABSTRACT FROM AUTHOR]

Due to their availability on commercial smartphones, WiFi, Bluetooth, and magnetometer are commonly utilized for indoor localization as indoor spaces are GPS deprived. Indoor localization falls into the category of data-intensive applications. In this domain, most of the recent solution approaches deploy machine learning (ML) and deep learning (DL) techniques on the data collected through the sensors. However, the publicly available benchmark datasets on indoor localization suffer from certain issues requiring complicated and customized data preprocessing techniques for each dataset for applying a common ML/DL technique. Thus, a fair comparison of the ML/DL methods for indoor localization datasets and hence to check for the generality of a solution that spans across different indoor regions become infeasible. In this comparative study, we have investigated three key challenges of fingerprint datasets that should be addressed for real-life localization applications, namely (i) repetitive site survey, (ii) device heterogeneity, and (iii) granularity and subregion-specific performance variation. To demonstrate how these attributes might impact localization performance, experimental analysis is performed using five benchmark datasets. The novelty of the work is that it not only highlights the challenges but also analyzes the feasibility of possible future directions to address these challenges through implementation results. Formulating the application of a generative adversarial network to address the issue of repetitive site surveys has been discussed with implementation results. [ABSTRACT FROM AUTHOR]

Compared to other well-known biometric technologies based on physiological traits (e.g., fingerprint, iris, and face), heart biometrics are more robust to presentation attacks and are particularly suitable for continuous/periodic recognition. Most studies on heart biometrics concern electrocardiogram (ECG) and photoplethysmogram (PPG). While the reported results are encouraging, to the best of our knowledge, no studies have been conducted on the interoperability between ECG and PPG biometrics. We present a novel method that is capable of performing single-domain and multiple-domain identity verifications for ECG and PPG signals, providing interoperability between the heterogeneous cardiac signals. Our method does not require the computation of any reference/fiducial point and uses a compact representation of the given signals. We propose MultiCardioNet, a novel Siamese neural network trained by using an ad hoc learning algorithm. MultiCardioNet computes a similarity score between two spectrogram-based representations of cardiac signals. Our learning algorithm iteratively computes a balanced subset of genuine and impostor pairs during the training epochs. We performed experiments on a dataset containing 1,008 pairs of ECG and PPG samples, obtaining accuracy comparable to that of the state-of-the-art methods for single-domain scenarios and demonstrating only a relatively small performance decrease in the multiple-domain scenario. [Display omitted] • The first study on the interoperability between ECG and PPG biometrics, based on Siamese CNNs. • Novel CNNs for processing two-dimensional representations of PPG signals. • Ad-hoc algorithm for training Siamese CNNs, suitable for small-size datasets. [ABSTRACT FROM AUTHOR]

Machine learning has gained attention due to its ongoing advancements and diverse applications. Within the field of homogeneous catalysis, a prominent area of research in machine learning revolves around predicting reaction yield in Pd‐catalyzed Buchwald–Hartwig amination reactions. This study sought to determine the optimal descriptors for representing the both structural and physical information associated with the reaction, particularly focusing on product details. To achieve this, we assessed the utilization of product extended molecular fingerprints (PEMF) and selected physical parameters (SPP). The utilization of a random forest model incorporating these descriptors yielded promising results in the prediction of reaction yields in Pd‐catalyzed Buchwald–Hartwig amination reactions. The model achieved an impressive R2 value of 0.943. Data preprocessing on PEMF and sorting preprocessing on physical parameters resulted in a significant reduction in data size to 259 bits PEMF+2 SPPs per prediction, much less than the two previous random forest models which utilized 480 physical parameters and 21,073 bits molecular fingerprints. Although establishing definitive correlations between SPPs and reaction yield presented challenges, our findings indicate that the presence of heavier atoms in the aryl halides may have a beneficial impact within the examined Pd‐catalyzed Buchwald–Hartwig amination reactions, as compared to their analogues. [ABSTRACT FROM AUTHOR]

Nowadays, an efficient and robust virtual screening procedure is crucial in the drug discovery process, especially when performed on large and chemically diverse databases. Virtual screening methods, like molecular docking and classic QSAR models, are limited in their ability to handle vast numbers of compounds and to learn from scarce data, respectively. In this study, we introduce a universal methodology that uses a machine learning-based approach to predict docking scores without the need for time-consuming molecular docking procedures. The developed protocol yielded 1000 times faster binding energy predictions than classical docking-based screening. The proposed predictive model learns from docking results, allowing users to choose their preferred docking software without relying on insufficient and incoherent experimental activity data. The methodology described employs multiple types of molecular fingerprints and descriptors to construct an ensemble model that further reduces prediction errors and is capable of delivering highly precise docking score values for monoamine oxidase ligands, enabling faster identification of promising compounds. An extensive pharmacophore-constrained screening of the ZINC database resulted in a selection of 24 compounds that were synthesized and evaluated for their biological activity. A preliminary screen discovered weak inhibitors of MAO-A with a percentage efficiency index close to a known drug at the lowest tested concentration. The approach presented here can be successfully applied to other biological targets as target-specific knowledge is not incorporated at the screening phase. [ABSTRACT FROM AUTHOR]

With the rapid development of the Internet of Things (IoT) technology, the number and types of devices accessing the Internet are increasing, leading to increased network security problems such as hacker attacks and botnets. Usually, these attacks are related to the type of device, and the risk can be effectively reduced if the type of network device can be efficiently identified and controlled. The traditional network device identification method uses active detection technology to obtain information about the device and match it with a manually defined fingerprint database to achieve network device identification. This method impacts the smoothness of the network and requires the manual establishment of fingerprint libraries, which imposes a large labor cost but only achieves a low identification efficiency. The traditional machine learning method only considers the information of individual packets; it does not consider the timing relationship between packets, and the recognition effect is poor. Based on the above research, in this paper, we considered the packet temporal relationship, proposed the TCN model of the Inception structure, extracted the packet temporal relationship, and designed a multi-head self-attention mechanism to fuse the features to generate device fingerprints for device identification. Experiments were conducted on the publicly available UNSW dataset, and the results showed that this method achieved notable improvements compared to the traditional machine learning method, with F1 reaching 96.76%. [ABSTRACT FROM AUTHOR]

Adverse drug reactions have, in some cases, devastating effects on the population, so it is very important to develop decision support systems based on artificial intelligence to predict such situations having a preventive effect with an important positive impact on population health. This research aims to study the influence of features, which are different characteristics used to represent chemical drug compounds, on the prediction of adverse reactions. A chemical fingerprint is defined as being a distinctive characteristic or pattern that indicates the presence of a certain molecule and a molecular descriptor that is a mathematical representation of a molecule's properties, which describes the chemical information of the molecule. Based on a comprehensive study of the state-of-the-art artificial intelligence algorithms such as Random Forest (RF), Gradient Boosted Trees (GBT), Feedforward Neural Network (FFN) and Convolutional Neural Network (CNN) are chosen, with which prediction models were created with different approaches to features such as the use of fingerprints, the use of molecular descriptors or the combination of fingerprints and molecular descriptors in order to obtain improved results. The algorithms specified above were selected since they are powerful and can capture complex patterns along the given features. The experimental evaluations consisted of identifying four effective fingerprints, namely: Pattern, Torsion, Atom Pair and Morgan, creating a machine learning model for each fingerprint type and using them as input data features, observing the results, after which two types of molecular descriptors were chosen, namely, molecular weight and number of valence electrons, we combined them as a feature of the input data, created machine learning models which we trained and observed the results, then, the last experiment, consisted of combining the molecular descriptors and the fingerprints, creating the machine learning models, training and observing the results. According to the experimental evaluation results RF and GBT had better results, with significant differences of even 20% in accuracy in certain cases, such as using molecular descriptors for Hepatobiliary disorders, than FNN and CNN. Another conclusion that can be formulated is the fact that the fingerprints had the greatest impact on the performance increase of the models, so the use of fingerprints and molecular descriptors together as features gave the best results among all the performed tests. [ABSTRACT FROM AUTHOR]

Siamese networks, representing a novel class of neural networks, consist of two identical subnetworks sharing weights but receiving different inputs. Here we present a similarity-based pairing method for generating compound pairs to train Siamese neural networks for regression tasks. In comparison with the conventional exhaustive pairing, it reduces the algorithm complexity from O(n2) to O(n). It also results in a better prediction performance consistently on the three physicochemical datasets, using a multilayer perceptron with the circular fingerprint as a proof of concept. We further include into a Siamese neural network the transformer-based Chemformer, which extracts task-specific features from the simplified molecular-input line-entry system representation of compounds. Additionally, we propose a means to measure the prediction uncertainty by utilizing the variance in predictions from a set of reference compounds. Our results demonstrate that the high prediction accuracy correlates with the high confidence. Finally, we investigate implications of the similarity property principle in machine learning. [ABSTRACT FROM AUTHOR]

Fingerprints are widely used as effective personal authentication systems, because they constitute unique, robust, and risk-free evidence. Fingerprinting techniques refer to biometric procedures used for identifying individuals based on their physical characteristics. A fingerprint image contains ridges and valleys forming a directionally-oriented pattern. The robustness of the fingerprint authentication technique determines the quality of the fingerprint image. This study proposed an intelligent 12-layered Convolutional Neural Network (CNN) model using Deep learning (DL) for gender determination based on fingerprints. Further, the study compared the performance of this model to existing state-of-the-art methods. The primary goal of this study was to reduce the number of comparisons within a large database obtained from automatic fingerprint recognition systems. The classification process was found to be swifter and more accurate when analysis of the DL algorithm was performed. With reference to the criteria of precision, recall, and accuracy evaluation during classification, this proposed 12-layered CNN model outperformed the Residual Neural Network with 50 Layers (ResNet-50) and Dense Convolutional Network with 201 Layers (DenseNet-201) models. The accuracies obtained were 97.0%, 95.8%, 98.0%, and 96.8% for female-left, female-right, male-left, and male-right classes respectively, while achieving an overall accuracy of 94.0%. [ABSTRACT FROM AUTHOR]

Aroma is a fundamental property of coffee and is modulated by volatile compounds (VCs). The VC fingerprint based on untargeted SPME-GC/MS can be used for coffee origin prediction. Coffee of Indonesian origin has been attracting considerable research interest in recent years. However, data analysis of the untargeted VC analysis remains one of the greatest challenges for the prediction of origins. Therefore, this study aimed to investigate the ability of untargeted SPME-GC/MS-based volatile fingerprinting to predict the origins of Indonesian coffee using machine learning (ML) approaches. Indonesian coffee samples from economically precious origins were studied. An SPME arrow was used to extract the VCs from the coffee headspace with optimised temperatures coupled to a GC/MS system. Several machine learning models were compared to obtain the most accurate origin prediction. This study found that adsorption at 70 °C for 10 min extracted the most reliable VCs for coffee origin prediction. The untargeted headspace-SPME volatile fingerprint of coffee employing 200 samples identified 224 features out of 656 detected signals. In the exploratory dataset, RF and PLS-DA models are comparable in predicting Indonesian coffee origins with accuracies of 97% and 95.2%, respectively. They also reached an AUC of 100% and 95.8% in the validation dataset, respectively. Furthermore, both models indicated promising results in selecting the important features. These features illustrate a clear classification in the visualisation using unsupervised models. Overall, the results of the study demonstrate the reliability of the current workflow for the predictive modelling of Indonesian coffees. This study contributes to the advancement of coffee origin prediction and classification for further authentication. [ABSTRACT FROM AUTHOR]

Questioned document examination aims to assess if a document of interest has been forged. Spectroscopy-based methods are the gold standard for this type of evaluation. In the past 15 years, Matrix-Assisted Laser Desorption Ionisation–Mass Spectrometry Imaging (MALDI-MSI) has emerged as a powerful analytical tool for the examination of finger marks, blood, and hair. Therefore, this study intended to explore the possibility of expanding the forensic versatility of this technique through its application to questioned documents. Specifically, a combination of MALDI-MSI and chemometric approaches was investigated for the differentiation of seven gel pens, through their ink composition, over 44 days to assess: (i) the ability of MALDI MSI to detect and image ink chemical composition and (ii) the robustness of the combined approach for the classification of different pens over time. The training data were modelled using elastic net logistic regression to obtain probabilities for each pen class and assess the time effect on the ink. This strategy led the classification model to yield predictions matching the ground truth. This model was validated using signatures generated by different pens (blind to the analyst), yielding a 100% accuracy in machine learning cross-validation. These data indicate that the coupling of MALDI-MSI with machine learning was robust for ink discrimination within the dataset and conditions investigated, which justifies further studies, including that of confounders such as paper brands and environmental factors. [ABSTRACT FROM AUTHOR]

Biometrics, e.g., fingerprints, the iris, and the face, have been widely used to authenticate individuals. However, most biometrics are not cancellable, i.e., once these traditional biometrics are cloned or stolen, they cannot be replaced easily. Unlike traditional biometrics, brain biometrics are extremely difficult to clone or forge due to the natural randomness across different individuals, which makes them an ideal option for identity authentication. Most existing brain biometrics are based on an electroencephalogram (EEG), which typically demonstrates unstable performance due to the low signal-to-noise ratio (SNR). Thus, in this paper, we propose the use of intracortical brain signals, which have higher resolution and SNR, to realize the construction of a high-performance brain biometric. Significantly, this is the first study to investigate the features of intracortical brain signals for identification. Specifically, several features based on local field potential are computed for identification, and their performance is compared with different machine learning algorithms. The results show that frequency domain features and time-frequency domain features are excellent for intra-day and inter-day identification. Furthermore, the energy features perform best among all features with 98% intra-day and 93% inter-day identification accuracy, which demonstrates the great potential of intracraial brain signals to be biometrics. This paper may serve as a guidance for future intracranial brain researches and the development of more reliable and high-performance brain biometrics. [ABSTRACT FROM AUTHOR]

Vascular endothelial growth factors (VEGFs) are specific cytokines involved in angiogenesis and do so via binding to vascular endothelial growth factor receptors (VEGFRs), a type of receptor tyrosine kinase. VEGFs are reported to facilitate angiogenesis in physiological (embryogenesis) and pathological (tumor) conditions. The overexpression of VEGFs and consequently VEGFRs is reported in tumorigenic conditions. Several VEGFR inhibitors currently used as anticancer drugs to prevent angiogenesis are sunitinib, sorafenib, etc. To identify new potential candidates as VEGFR inhibitors, a classification study using a large and diverse dataset of VEGFR inhibitors from the BindingDB database has been conducted. The KNIME platform was used to calculate molecular and fingerprint‐based descriptors and several classification algorithms viz. linear regression (LR), k‐nearest neighbor (kNN), decision tree (DT), random forest (RF), and gradient boosted tree (GBT) were employed to build the classification model. The model performance was evaluated by accuracy, precision, recall, and F1 score of the test set. The best LR, kNN, DT, RF, and GBT classifiers had the F1 score of 0.81, 0.87, 0.82, 0.87, and 0.87, respectively. The assorted 5120 VEGFR inhibitors were clustered into 10 subsets, and the structural features of each subset were assessed along with the identification of significant fragments in active and inactive compounds. The automated classifier model developed using the KNIME platform could serve as an important platform for screening and designing molecules as VEGFR inhibitors. [ABSTRACT FROM AUTHOR]

Identifying an individual from accelerometry data collected during walking without reliance on step-cycle detection has not been achieved with high accuracy. We propose an open-source reproducible method to: (1) create a unique, person-specific "walking fingerprint" from a sample of un-landmarked high-resolution data collected by a wrist-worn accelerometer; and (2) predict who an individual is from their walking fingerprint. Accelerometry data were collected during walking from 32 individuals (23–52 y.o., 19 females) for at least 380 s each. For this study's purpose, data are not landmarked, nor synchronized. Individual walking fingerprints were created by: (1) partitioning the accelerometer time series in adjacent, non-overlapping one-second intervals; (2) transforming all one-second interval data for a given individual into a three-dimensional (3D) image obtained by plotting each one-second interval time series by the lagged time series for a series of lags; (3) partitioning these resulting participant-specific 3D images into a grid of cells; and (4) identifying the combinations of cells (areas in the 3D image) that best predict the individual. For every participant, the first 200 s of data were used as training and the last 180 s as testing. This approach does not use segmentation methods for individual strides, which reduces dependence on complementary algorithms and increases its generalizability. The method correctly identified 100 % of the participants in the test data and highlighted unique features of walking that characterize the individuals. Predicting the identity of an individual from their walking pattern has immediate implications that can complement or replace those of actual fingerprinting, voice, and image recognition. Furthermore, as walking may change with age or disease burden, individual walking fingerprints may be used as biomarkers of change in health status with potential clinical and epidemiologic implications. • No methods exist to identify individuals from high-resolution, un-landmarked, accelerometry-derived data from walking. • Time-series accelerometry data from walking can be transformed into a 3D image. • Predictors can be derived from these 3D images and used in classification algorithms. • Multivariate logistic regression using these predictors correctly identifies 100 % of subjects in a study of 32 individuals. [ABSTRACT FROM AUTHOR]

The number of Cyber-Physical Systems (CPS) available in industrial environments is growing mainly due to the evolution of the Internet-of-Things (IoT) paradigm. In such a context, radio frequency spectrum sensing in industrial scenarios is one of the most interesting applications of CPS due to the scarcity of the spectrum. Despite the benefits of operational platforms, IoT spectrum sensors are vulnerable to heterogeneous malware. The usage of behavioral fingerprinting and machine learning has shown merit in detecting cyberattacks. Still, there exist challenges in terms of (i) designing, deploying, and evaluating ML-based fingerprinting solutions able to detect malware attacks affecting real IoT spectrum sensors, (ii) analyzing the suitability of kernel events to create stable and precise fingerprints of spectrum sensors, and (iii) detecting recent malware samples affecting real IoT spectrum sensors of crowdsensing platforms. Thus, this work presents a detection framework that applies device behavioral fingerprinting and machine learning to detect anomalies and classify different botnets, rootkits, backdoors, ransomware and cryptojackers affecting real IoT spectrum sensors. Kernel events from CPU, memory, network, file system, scheduler, drivers, and random number generation have been analyzed, selected, and monitored to create device behavioral fingerprints. During testing, an IoT spectrum sensor of the ElectroSense platform has been infected with ten recent malware samples (two botnets, three rootkits, three backdoors, one ransomware, and one cryptojacker) to measure the detection performance of the framework in two different network configurations. Both supervised and semi-supervised approaches provided promising results when detecting and classifying malicious behaviors from the eight previous malware and seven normal behaviors. In particular, the framework obtained 0.88–0.90 true positive rate when detecting the previous malicious behaviors as unseen or zero-day attacks and 0.94–0.96 F1-score when classifying them. [ABSTRACT FROM AUTHOR]

The fingerprint comes to be the most popular and utilized biometric for identifying persons owing to its bio-invariant characteristic, precision, as well as easy acquisition. A sub-system of an identification system is the classification stage in order to diminish the penetration rate and computational complexity. Actually, there are many formal investigations regarding techniques by exploiting convolutional neural networks (CNN) together with fingerprints images, which have superior performance metrics at the cost of large training times even employing high performance computing, which is not feasible in the standard word. In our manuscript, researches about identify and classify fingerprint databases by recurring to extreme learning machines (ELM) will be extensively reported and discussed for the first time. The diverse methodologies (ELM plus feature extractors) given by the authors will be studied and contrasted considering performance analysis. Consequently, academic papers with diverse version of ELMs are developed to observe the pros and cons that they exhibit with each other and to probe how they may help for minimizing the penetration rate of fingerprint databases. In fact, this issue is very relevant because by enhancing the penetration rate means shorting search times and computational complexity in fingerprints. [ABSTRACT FROM AUTHOR]

A novel fingerprinting platform for multiplex detection of flavor molecules in Baijiu was developed by using a surface-enhanced Raman scattering (SERS) nanosensor array in combination with machine learning. The SERS sensors were constructed by core–shell Fe3O4@Ag nanoparticles modified with molecules carrying end-groups of hydroxyl, pyridyl, methyl, and amino, respectively, which interacted with flavors and led to changes in the sensors' spectra. All the Raman spectra acquired from the nanosensor array contacting with the sample were concatenated into a single SERS super-spectrum, representing the flavor fingerprint which was recognized through machine learning. Principal component analysis, support vector machine, and partial least squares were utilized to build classification and quantitation models for predictive analyses. The SERS nanosensor array was successfully used for fingerprinting ten typical flavors in Baijiu including four esters, three alcohols, and three acids, with an accuracy of 100%, linear detection ranges over two orders of magnitude, and limits of detection ranging from 3.45 × 10−3 mg/L of phenylethyl acetate to 1.21 × 10−2 mg/L of ethyl hexanoate. It was also demonstrated that satisfactory accuracies (recoveries) ranging from 96.2 to 104% and relative standard deviations ranging from 0.65 to 2.78% were obtained for the simultaneous quantification of 3-methylbutyl acetate and phenylethyl acetate in eighteen Baijiu samples of three flavor types including sauce flavor, strong flavor, and light flavor. Compared with the existing detection techniques, this chemical fingerprinting platform is easy to use, highly sensitive, and can perform multiplex detection, which has great potential for practical applications. [ABSTRACT FROM AUTHOR]

Structural aberrations (SA) have been shown to play an essential role in the occurrence and development of cancer. SAs are typically characterized by copy number alteration (CNA) dose and distortion length. Although sequencing techniques and analytical methods have facilitated the identification and cataloging of somatic CNAs, there are no effective methods to quantify SA considering the amplitude, location, and neighborhood of each nucleotide in each fragment. Therefore, a new SA index based on dynamic time warping is proposed. The SA index analysed 22448 samples of 35 types/subtypes of cancers. Most types had significant differences in SA levels ranging between 12p and 20q. This suggests that genes or inter-gene regions may warrant greater attention, as they can be used to distinguish between different types of cancers and become targets for specific treatments. SA indexes were then used to quantify the differences between cancers. Additionally, SA fingerprints were identified for every cancer type. Kidney chromophobe, adrenocortical carcinoma, and ovarian serous cystadenocarcinoma are the three severest types with structural aberrations caused by cancer, while thyroid carcinoma is the least. Our research provides new possibilities for the better utilization of chromosomal instability for further exploiting cancer aneuploidy, thus improving cancer therapy. [ABSTRACT FROM AUTHOR]

This review paper analyses the utilization of machine learning algorithm in facial identity verification of computer-based assessments. Through the use of facial recognition, this project aims to increase security by letting a customer know who is actually using the system. Only authorized customers are able to enter the framework. Neural learning approaches for face recognition can be used for feature extraction and training modules. Furthermore, a lot of people use similar methods to extract information from photographs of people. Certain detection systems can perform full body scans, as well as iris and finger print recognition. The purpose of these systems' deployment is safety and security. In this study, we examine many facial recognitions machine learning techniques. Four supervised machine-learning classifiers for face recognition are taken into consideration: Support Vector Machine (SVM), 1-nearest neighbor (1-NN), Principal Component Analysis (PCA), and Linear Discriminant Analysis (LDA). The accuracy with which different categorization schemes can recognize a face is another way in which their efficacy is proven and evaluated. The process of identifying faces of individuals whose photos are kept in databases and made available as datasets is known as face recognition. Numerous tests carried out using these datasets. Which machine learning method is the best in terms of picture detection accuracy is made evident by the comparative study. Even if there are other very successful identification techniques, face recognition has remained a prominent area of study interest because it is a simple and non-intrusive way for people to identify themselves. The results of this study could be helpful in determining the best machine-learning method to improve the accuracy of facial recognition. [ABSTRACT FROM AUTHOR]

Cytochrome P450 (CYP) enzymes are involved in the metabolism of approximately 75% of marketed drugs. Inhibition of the major drug‐metabolizing P450s could alter drug metabolism and lead to undesirable drug–drug interactions. Therefore, it is of great significance to explore the inhibition of P450s in drug discovery. Currently, machine learning including deep learning algorithms has been widely used for constructing in silico models for the prediction of P450 inhibition. These models exhibited varying predictive performance depending on the use of machine learning algorithms and molecular representations. This leads to the difficulty in the selection of appropriate models for practical use. In this study, we systematically evaluated the conventional machine learning and deep learning models for three major P450 enzymes, CYP3A4, CYP2D6, and CYP2C9 from several perspectives, such as algorithms, molecular representation, and data partitioning strategies. Our results showed that the XGBoost and CatBoost algorithms coupled with the combined fingerprint/physicochemical descriptor features exhibited the best performance with Area Under Curve (AUC) of 0.92, while the deep learning models were generally inferior to the conventional machine learning models (average AUC reached 0.89) on the same test sets. We also found that data volume and sampling strategy had a minor effect on model performance. We anticipate that these results are helpful for the selection of molecular representations and machine learning/deep learning algorithms in the P450 model construction and the future model development of P450 inhibition. Inhibition of the major drug‐metabolizing cytochrome P450 enzymes could lead to serious adverse reactions. Currently, a variety of machine learning models have been developed for the prediction of P450 inhibitors, but a detailed comparison is lacking. Different algorithms and representations were used to build the predictive models, and their performance was compared here. [ABSTRACT FROM AUTHOR]

Infrared spectroscopy (IR) is a widely used technique enabling to identify specific functional groups in the molecule of interest based on their characteristic vibrational modes or the presence of a specific adsorption site based on the characteristic vibrational mode of an adsorbed probe molecule. The interpretation of an IR spectrum is generally carried out within a fingerprint paradigm by comparing the observed spectral features with the features of known references or theoretical calculations. This work demonstrates a method for extracting quantitative structural information beyond this approach by application of machine learning (ML) algorithms. Taking palladium hydride formation as an example, Pd‐H pressure‐composition isotherms are reconstructed using IR data collected in situ in diffuse reflectance using CO molecule as a probe. To the best of the knowledge, this is the first example of the determination of continuous structural descriptors (such as interatomic distance and stoichiometric coefficient) from the fine structure of vibrational spectra, which opens new possibilities of using IR spectra for structural analysis. [ABSTRACT FROM AUTHOR]

The skin surface is an important sample source that the metabolomics community has only just begun to explore. Alterations in sebum, the lipid-rich mixture coating the skin surface, correlate with age, sex, ethnicity, diet, exercise, and disease state, making the skin surface an ideal sample source for future noninvasive biomarker exploration, disease diagnosis, and forensic investigation. The potential of sebum sampling has been realized primarily via electrospray ionization mass spectrometry (ESI-MS), an ideal approach to assess the skin surface lipidome. However, a better understanding of sebum collection and subsequent ESI-MS analysis is required before skin surface sampling can be implemented in routine analyses. Challenges include ambiguity in definitive lipid identification, inherent biological variability in sebum production, and methodological, technical variability in analyses. To overcome these obstacles, avoid common pitfalls, and achieve reproducible, robust outcomes, every portion of the workflow—from sample collection to data analysis—should be carefully considered with the specific application in mind. This review details current practices in sebum sampling, sample preparation, ESI-MS data acquisition, and data analysis, and it provides important considerations in acquiring meaningful lipidomic datasets from the skin surface. Forensic researchers investigating sebum as a means for suspect elimination in lieu of adequate fingerprint ridge detail or database matches, as well as clinical researchers interested in noninvasive biomarker exploration, disease diagnosis, and treatment monitoring, can use this review as a guide for developing methods of best-practice. [ABSTRACT FROM AUTHOR]

Every year, millions of students graduate, making it difficult to track and process their data. The One Point Student Verification Machine learning Model (OSVMLM) collects rich data on students, including biometric information, secondary education details, list of degrees completed, and detailed information on their last or current degree. In addition, the student's fingerprint is also collected and stored. The government's Aadhar database is used to verify the authenticity of the student. To ease the recruiter, the details collected by OSVMLM were assigned with ranks, information will be sent to the candidates autonomously based on the current job description. This makes it easy for students to identify and apply for available positions. OSVMLM application is built on Ionic, MongoDB, NodeJS and Gradient boosting machines (GBMs). OSVMLM is immensely useful for recruiters, as it makes data collection and processing effortless. This increases the visibility of various jobs open in different domains and suggests suitable job description to the job seekers. The key difference between OSVMLM and other platforms is that it is not limited to institutions and is accessible to all. This saves a lot of time for recruiters during initial screening. This proposed OSVMLM was validated with the real-time students dataset and achieved 91% prediction accuracy. [ABSTRACT FROM AUTHOR]

This study focuses on enhancing office security through a smart door system, designed to protect sensitive documents and critical data. Emphasizing exclusive access for authorized personnel, the system integrates advanced biometric authentication, predominantly facial recognition. The project's aim is to optimize face recognition using convolutional neural network (CNN) techniques, identifying the best preprocessing methods and hyperparameter settings. A significant aspect of the research involves developing a smart door system with remote authentication and control capabilities via internet connectivity. Employing transfer learning with MobileNet V2, the study presents a compact model tailored for the Raspberry Pi platform. The model utilizes a dataset with five facial recognition classes and an additional class for unknown faces, ensuring a diverse representation. The trained model achieved a high accuracy (0.9729) and low loss (0.09). System evaluation revealed an overall accuracy of 0.96, perfect recall (1.00), and a precision of 0.897. These results demonstrate the system's efficacy in secure access control, making it a viable solution for contemporary office environments. [ABSTRACT FROM AUTHOR]

Histone deacetylase 6 (HDAC6) is a promising drug target for the treatment of human diseases such as cancer, neurodegenerative diseases (in particular, Alzheimer's disease), and multiple sclerosis. Considerable attention is paid to the development of selective non-toxic HDAC6 inhibitors. To this end, we successfully form a set of 3854 compounds and proposed adequate regression QSAR models for HDAC6 inhibitors. The models have been developed using the PubChem, Klekota-Roth, 2D atom pair fingerprints, and RDkit descriptors and the gradient boosting, support vector machines, neural network, and k-nearest neighbours methods. The models are integrated into the developed HT_PREDICT application, which is freely available at . In vitro studies have confirmed the predictive ability of the proposed QSAR models integrated into the HT_PREDICT web application. In addition, the virtual screening performed with the HT_PREDICT web application allowed us to propose two promising inhibitors for further investigations. [ABSTRACT FROM AUTHOR]

Radio Frequency Fingerprinting (RFF) refers to the technique for identifying and classifying wireless devices on the basis of their physical characteristics, which appear in the digital signal transmitted in space. Small differences in the radio frequency front-end of the wireless devices are generated across the same wireless device model during the implementation and manufacturing process. These differences create small variations in the transmitted signal, even if the wireless device is still compliant with the wireless standard. By using data analysis and machine-learning algorithms, it is possible to classify different electronic devices on the basis of these variations. This technique has been well proven in the literature, but research is continuing to improve the classification performance, robustness to noise, and computing efficiency. Recently, Deep Learning (DL) has been applied to RFF with considerable success. In particular, the combination of time-frequency representations and Convolutional Neural Networks (CNN) has been particularly effective, but this comes at a great computational cost because of the size of the time-frequency representation and the computing time of CNN. This problem is particularly challenging for wireless standards, where the data to be analyzed is extensive (e.g., long preambles) as in the case of the LoRa (Long Range) wireless standard. This paper proposes a novel approach where two pre-processing steps are adopted to (1) improve the classification performance and (2) to decrease the computing time. The steps are based on the application of Variational Mode Decomposition (VMD) where (in opposition to the known literature) the residual of the VMD application is used instead of the extracted modes. The concept is to remove the modes, which are common among the LoRa devices, and keep with the residuals the unique intrinsic features, which are related to the fingerprints. Then, the spectrogram is applied to the residual component. Even after this step, the computing complexity of applying CNN to the spectrogram is high. This paper proposes a novel step where only segments of the spectrogram are used as input to CNN. The segments are selected using a machine-learning approach applied to the features extracted from the spectrogram using the Local Binary Pattern (LBP). The approach is applied to a recent LoRa radio frequency fingerprinting public data set, where it is shown to significantly outperform the baseline approach based on the full use of the spectrogram of the original signal in terms of both classification performance and computing complexity. [ABSTRACT FROM AUTHOR]

In the realm of today's networking technologies, user localization has been a formidable challenge for recent applications. There are different approaches in pursuit of heightened position detection of an end-user with the help of GPS, Wi-Fi fingerprint and 5G equipment. However, these approaches require both deployment and maintenance costs because of equipment establishment for position tracking. Moreover, they are not capable of minimizing the localization error, especially for indoor scenarios to track the indoor position of an end-user. Hence, there is an urgent need to delve deeper into innovative approaches to drive further advancements in user localization. In response, Machine Learning (ML) approaches have recently been widely adapted to predict the localization of end-users with minimum error. More specifically, average localization error (ALE) of an end-user can be predicted in a cost-effective way by using proper data and ML methods. For this purpose, we have investigated different ML approaches to get an accurate ALE prediction scheme for 5G networks with mobile end-users. Accordingly, an existing dataset is utilized to generate localization data of end-users in which the ALE is directly calculated by Received Signal Strength Indicator. Moreover, three different normalization approaches are applied for the overarching goal of increased data quality. Consequently, six different ML algorithms, including Linear regression, support vector machine with three different kernels, Gaussian process, and ensemble least-squares boosting (LSBoost) are evaluated with respect to a set of evaluation criteria including R, R2, RMSE, and MAE. The evaluation outcomes emphasize that ensemble LSBoost method, in the context of localization prediction, outperforms the other approaches and is sufficient to yield a viable learning strategy for ALE prediction. [ABSTRACT FROM AUTHOR]