Discriminant Model Research Articles

The inclusive data governance models for algorithms – a dream of the already convinced or a realistic way of action?

ABSTRACT The article presents the implications of personal data governance models for algorithmic discrimination. It reflects on the significance of engaging with data subjects to prevent or mitigate discrimination by comparing the effectiveness of the top-down and bottom-up approaches to data governance. The paper goes beyond the option to either decline or agree to terms of service while entering online platforms and argues that bottom-up data governance models of data trusts and information fiduciaries can be more inclusive than the top-down data governance.

Omega‐3 Marine Fatty Acid Supplementation to Healthy Subjects Interacts With Moderate Physical Activity to Provide a Cardiovascular Healthier Lipoprotein Subclass Profile

ABSTRACTThis work investigates the impact of marine omega‐3 and physical activity and their interaction on cardiometabolic health as expressed by the serum lipoprotein profile. Using an experimental design that allows for the possibility of interaction, we performed a 6‐week intervention on 44 middle‐aged women living in Western Norway. The women were randomly divided into four groups: one control group with no intervention, a second group performing sessions of moderate intensity three times per week, a third group taking daily supplements of omega‐3 marine fatty acids, and a fourth group combining the interventions for Groups 2 and 3. The difference in the lipoprotein profiles after the intervention from baseline were assessed for statistical significance by comparing groups 2, 3 and 4 with Group 1 using two‐tailed t‐test corrected for multiple testing and selectivity ratios calculated from the discriminatory component in validated partial least squares discriminant models. The results from the univariate and multivariate analyses were qualitatively equivalent: Only the women combining moderate physical activity and omega‐3 supplementation, revealed statistically significant differences in their lipoprotein profile compared to the nonintervention control group. The pattern of change in the lipoprotein profile is associated with improved cardiometabolic health. Use of the design matrix to predict this pattern revealed that the interaction between omega‐3 supplementation and physical activity played a major role in inducing this change. The recognition of the influence of this interaction may be a step towards resolving the long‐lasting debate of the role played by omega‐3 for preventing cardiovascular unhealth.

Chemometric Discrimination of the Geographical Origin of Rheum tanguticum by Stable Isotope Analysis

Rheum tanguticum is one of the primary rhubarb species used for food and medicinal purposes, and it has recently been gaining more attention and recognition. This research represents the first attempt to use stable isotopes and elemental analysis via IRMS to identify the geographical origin of Rheum tanguticum. A grand total of 190 rhubarb samples were gathered from 38 locations spread throughout the provinces of Gansu, Sichuan, and Qinghai in China. The carbon content showed a decreasing trend in the order of Qinghai, followed by Sichuan, and then Gansu. Nitrogen content was notably higher, with Qinghai and Sichuan displaying similar levels, while Gansu had the lowest nitrogen levels. Significant differences were noted in the δ13C (−28.9 to −26.5‰), δ15N (2.6 to 5.6‰), δ2H (−120.0 to −89.3‰), and δ18O (16.0‰ to 18.8‰) isotopes among the various rhubarb cultivation areas. A significant negative correlation was found between %C and both longitude and humidity. Additionally, δ13C and δ15N isotopes were negatively correlated with longitude, and δ15N showed a negative correlation with humidity as well. δ2H and δ18O isotopes exhibited a strong positive correlation with latitude, while significant negative correlations were observed between δ2H and δ18O isotopes and temperature, precipitation, and humidity. The LDA, PLS-DA, and k-NN models all exhibited strong classification performance in both the training and validation sets, achieving accuracy rates between 82.1% and 91.7%. The combination of stable isotopes, elemental analysis, and chemometrics provides a reliable and efficient discriminant model for accurately determining the geographical origin of R. tanguticum in different regions. In the future, the approach will aid in identifying the geographical origin and efficacy of rhubarb in other studies.

Sound Sensing: Generative and Discriminant Model-Based Approaches to Bolt Loosening Detection.

The detection of bolt looseness is crucial to ensure the integrity and safety of bolted connection structures. Percussion-based bolt looseness detection provides a simple and cost-effective approach. However, this method has some inherent shortcomings that limit its application. For example, it highly depends on the inspector's hearing and experience and is more easily affected by ambient noise. In this article, a whole set of signal processing procedures are proposed and a new kind of damage index vector is constructed to strengthen the reliability and robustness of this method. Firstly, a series of audio signal preprocessing algorithms including denoising, segmenting, and smooth filtering are performed in the raw audio signal. Then, the cumulative energy entropy (CEE) and mel frequency cepstrum coefficients (MFCCs) are utilized to extract damage index vectors, which are used as input vectors for generative and discriminative classifier models (Gaussian discriminant analysis and support vector machine), respectively. Finally, multiple repeated experiments are conducted to verify the effectiveness of the proposed method and its ability to detect the bolt looseness in terms of audio signal. The testing accuracy of the trained model approaches 90% and 96.7% under different combinations of torque levels, respectively.

Introducing the discrimination model for acceptance and commitment therapy supervision

AbstractThe Discrimination Model for Acceptance and Commitment Therapy Supervision represents a combined model that integrates six core processes of acceptance commitment therapy (ACT) with the multi‐dimensional structure of the Discrimination Model (DM). Based on empirical evidence of ACT and DM, respectively, we present guidelines and case vignettes to demonstrate the implementation of DMACTS in counselor supervision.

Public firms on an international border: A model of spatial price discrimination

This paper uniquely examines an international mixed oligopoly in a model of spatial price discrimination. It isolates the importance of the location of the border showing a variety of equilibria depending on the nationality and placement of the private rivals. While the presence of a public firm often improves domestic welfare, it need not. Moreover, a prisoner's dilemma can exist in which each country would benefit from the privatization of both public firms but neither country has a unilateral incentive to privatize. The implications are discussed.

Towards a Spectral Library of Medicinal and Aromatic Plant species (MAPs): Plant Discrimination and Wavelength Selection

The recognition and identification of Medicinal and Aromatic Plants species (MAPs) is a challenge for researchers and professionals in the field. To address this issue, this study aimed to examine the potential of using hyperspectral remote sensing, chemometrics and Machine Learning in discriminating between MAP species. Specifically, we tested the applicability of UV-Near-infrared Spectroscopy (UV-NIR), PCA and Partial Least-Squares Discriminant Analysis (PLS-DA) as tools for discrimination. The spectral signature data were obtained by UV-near infrared spectroscopy, covering a range of 325–1075 nm. Principal component analysis (PCA) was applied to reduce data dimensionality in order to minimize errors in PLS-DA model. This model was used to build a discriminant model based on the wavelengths obtained from PCA. The results show that it is effective to combine UV-NIR spectral features and PLS-DA to establish a discriminant model for accurately classifying the MAP species. The key wavelengths most involved in the discrimination were found in the NIR region and especially in ranges 516–548 and 628–1070 nm.

Non-Destructive Identification of Wool and Cashmere Fibers Based on Cascade Optimizations of Interval-Wavelength Selection Using NIR Spectroscopy

ABSTRACT Near-infrared (NIR) spectroscopy is an effective method for identifying wool and cashmere fibers, with high spectral data providing a wealth of information. However, a key issue is that the accuracy and robustness of subsequent estimates can be reduced by redundant and interfering wavelengths. For this reason, a novel interval-wavelength cascaded optimization method is proposed. Initially, the collected spectral data are preprocessed by standard normal variate transformation (SNV) to eliminate the scattering effect. Then, the backward interval partial least squares (BiPLS) algorithm is applied for the preliminary selection of spectral intervals, followed by the application of three different variable selection algorithms, competitive adaptive reweighted sampling (CARS), successive projection algorithm (SPA) and whale optimization algorithm (WOA), for secondary wavelength optimization, respectively. Finally, both support vector machine (SVM) and random forest (RF) discriminant models are built to identify the extracted subset of wavelengths. In the experimental stage, the cascade method BiPLS-WOA selects 36 wavelengths, in SVM, the accuracy of the validation set reaches 96.9%, and the area under the ROC curve (AUC) can reach 99.3%. The results demonstrate that the proposed method can eliminate redundant and collinear variables, thereby validating the effectiveness of distinguishing wool and cashmere fibers.

Robust generalized PCA for enhancing discriminability and recoverability

The dependency of low-dimensional embedding to principal component space seriously limits the effectiveness of existing robust principal component analysis (PCA) algorithms. Simply projecting the original sample coordinates onto orthogonal principal component directions may not effectively address various noise-corrupted scenarios, impairing both discriminability and recoverability. Our method addresses this issue through a generalized PCA (GPCA), which optimizes regression bias rather than sample mean, leading to more adaptable properties. And, we propose a robust GPCA model with joint loss and regularization based on the ℓ2,μ norm and ℓ2,ν norms, respectively. This approach not only mitigates sensitivity to outliers but also enhances feature extraction and selection flexibility. Additionally, we introduce a truncated and reweighted loss strategy, where truncation eliminates severely deviated outliers, and reweighting prioritizes the remaining samples. These innovations collectively improve the GPCA model’s performance. To solve the proposed model, we propose a non-greedy iterative algorithm and theoretically guarantee the convergence. Experimental results demonstrate that the proposed GPCA model outperforms the previous robust PCA models in both recoverability and discrimination.

Evaluation of 1064 nm Raman for meat and bone meal species discrimination: From raw form to chemical and physical components

The 1064 nm Raman spectroscopy technique was employed to characterize raw, bone fragments (BF), meat fragments (MF), lipid, and residue of meat and bone meal (MBM) samples. Employing three chemometric algorithms, species discrimination models were optimized for the aforementioned sample types, indicating that the presence of BF favored discriminant analysis for poultry samples, yielding a classification error of 0.000. The species discrimination model constructed using Raman spectra combined with PLS-DA algorithm for residue samples was found to be overall optimal, with classification error for four species samples all below 0.061. A detailed analysis of the BF related spectral variables contributing significantly to PLS-DA discriminant analysis was conducted. Finally, this study established the complementarity of Raman spectroscopy in species discrimination analysis from both a physical and chemical components perspective, laying the theoretical groundwork for the development of a high-precision 1064 nm Raman spectroscopic analysis method for MBM species discrimination.

Pattern recognition-based aptasensor array for discrimination of matrix effect

Nucleic acid aptamers have attracted considerable attention in recent years, particularly in the fields of disease diagnosis and treatment and biosensing. However, complex matrix effects have severely hindered their progress towards practical application. In this study, the novel strategy of a pattern recognition-based aptasensor array was introduced for target recognition based on the predictable selectivity capability of aptamers by circumventing the challenge of the interference of the matrix effect. As one proof of concept, we chose lactoferrin as the target protein, three selected aptamers with differential affinity as its recognition probes, eleven matrix samples of milk powder as the pattern discrimination model, and AuNPs color as the response signals. In the AuNPs colorimetric arrays, the lactoferrin-added milk powder matrix induced different degrees of protection on the surface of AuNPs particles due to different degrees of binding of lactoferrin to the three aptamer sequences and therefore exhibited differentiated red-to-blue color change along with salt-induced aggregation behaviors. The difference in color responses exhibited with and without lactoferrin addition was considered as the fingerprint pattern of the lactoferrin in each milk powder matrix, followed by the supervised machine learning analysis of linear discriminant analysis (LDA) for better orthogonality to complete the identification and prediction of the unknown samples. By this strategy, eleven kinds of milk powder at the added concentration of lactoferrin (50 nmol/L) presented a distinct response pattern and have been identified and classified with accuracies of 100 % via LDA patterns (three aptamers × eleven milk powder matrices × five replicates) with three canonical factors (91.2 %, 8.58 %, and 0.22 %). Furthermore, the accuracy of 22 unknown samples was 100 %, indicating the achievement in differentiating between different milk powder matrices and the identification of each matrix. The proposed aptasensor array complements the traditional “lock-and-key” single aptasensor and opens a new direction for developing sensitive sensing array systems circumventing complex matrix effects.

Constructing a predictive model based on peripheral blood signs to differentiate infectious mononucleosis from chronic active EBV infection.

To develop a prediction model based on peripheral blood signs to distinguish between infectious mononucleosis and chronic active EBV infection. Retrospective data was collected for 60 patients with IM (IM group) and 20 patients with CAEBV infection (CAEBV group) who were hospitalized and diagnosed at the General Hospital of Tianjin Medical University between December 2018 and September 2022. The analyses used were univariate and LASSO (least absolute shrinkage and selection operator) logistic regression. Univariate analyses revealed that both IM and CAEBV-infected patients displayed overlapping and intersecting clinical manifestations, such as fever, sore throat, enlarged lymph nodes, and enlargement of the liver and spleen, and that in contrast to inflammatory responses in peripheral blood, CAEBV-infected patients had more severe inflammatory responses. Nine biomarkers-HGB, lymphocyte count, percentage of lymphocytes, ALB, fibrinogen, CRP, IFN-, IL-6, and EBV-DNA load-were subsequently selected by LASSO logistic regression modeling to serve as discriminatory models. Our investigation offers a solid foundation for diagnosing IM and CAEBV infection using the LASSO logistic regression model based on the significance and availability of peripheral blood indicators. Infected patients with CAEBV require early medical attention.

Predicting the degree of rubber rupture damage using a GAN-enhanced Bayesian-optimized 1DCNN network

Influenced by factors such as temperature, aging, and overloading, rubber bearings may undergo rupture during prolonged usage, leading to severe consequences such as bearing failure and structural damage. To accurately assess the degree of rubber rupture damage, this study proposed a novel generative adversarial network (GAN)-enhanced Bayesian-optimized one-dimensional convolutional neural network (1DCNN) framework (GAN-BCNN). In the GAN-BCNN framework, GAN is used to enhance the proportion of damaged data in the database, while Bayesian optimization is utilized to optimize crucial hyperparameters in the 1DCNN network. The predictive results indicate that the proposed GAN-BCNN model accurately predicts the degree of rubber rupture damage, achieving high classification accuracies of 99.1, 95.5, 92.9, and 98.0% for the categories of no damage, slight damage, moderate damage, and severe damage, respectively. When compared to the linear discriminant analysis model and the BCNN model without GAN enhancement, the proposed GAN-BCNN model exhibits a significant improvement in predictive accuracy, with an increase in Accuracy values of 52.4 and 6.0%, respectively. These results indicate that the GAN-BCNN model holds promising practical applications, and the GAN algorithm contributes to improving the prediction accuracy of the model.

Toward the unification of generative and discriminative visual foundation model: a survey

Toward the unification of generative and discriminative visual foundation model: a survey

Integrating mineral elements and metabolite features to distinguish Lotus seeds from different geographic origins

The characteristics of lotus seeds (LS) are influenced by variety and environment. However, it remains unknown the difference of metabolites and elements of LS from different origins. In this study, an accurate quantification method (97–107 %) for 20 mineral elements in LS was developed, and a metabolomic method was established to identify a total of 323 metabolites in LS. Mineral composition analysis revealed significant variations in the mineral element contents among LS samples from seven geographical regions. LS were rich in potassium (14,710 mg/kg), manganese (67.19 mg/kg), with a low level of sodium (210 mg/kg). A total of 10 mineral elements and 117 metabolites (p < 0.05 and VIP > 1) were identified as the potential geographical markers of LS by integration analysis. The linear discriminant analysis model showed high prediction accuracy. This study provides strong experimental evidence to maintain the authenticity and quality of LS in the food industry.

AMTrack:Transformer tracking via action information and mix-frequency features

Nowadays, Transformer-based visual tracking algorithms have been developing quickly because of the self-attention mechanism of Transformer, which has the capability to model global information. Although the self-attention mechanism in Transformer can effectively capture long-range dependencies in feature space, they only use flattened two-dimensional features and are unable to capture long-range temporal dependencies. Furthermore, since the self-attention in Transformer functions as a low-pass filter, it picks up on low-frequency features of the target while ignoring high-frequency features. This research suggests a Transformer tracker based on action information and mix-frequency features (AMTrack) to address these problems. Specifically, to address the lack of temporal remote dependencies, we introduce the target action aware module and the target action offset module. The target action aware module sets up several pathways to extract spatio-temporal, channel, and motion feature independently. In contrast, the target action offset module computes the target’s offset information by computing relative feature maps. Furthermore, in order to address the imbalance between high and low frequency features, we propose a mix-frequency attention and multi-frequency self-attention convolutional block. The mix-frequency attention uses high-frequency features within partitioned local windows as input for the high-frequency branch and average-pooled low-frequency features as the input for the low-frequency branch, calculating attention scores in both branches respectively. The multi-frequency self-attention convolutional block uses self-attention to capture low-frequency features and convolution to capture high-frequency features. Extensive experiments are carried out on eight challenging tracking datasets (e.g., OTB100 (Object Tracking Benchmark 100), NFS (Need For Speed), UAV123 (Unmanned Aerial Vehicles 123), TC128 (Temple Color 128), VOT2018 (Visual Object Tracking 2018), LaSOT (Large-scale Single Object Tracking), TrackingNet (Tracking Network), GOT-10k (Generic Object Tracking-10k)), and the experimental results show that our tracker achieves excellent tracking performance when compared with several state-of-the-art tracking algorithms. The experimental results show that on LaSOT, the success rates AUC (Area Under Curve), PNorm, and P reach 65.8%, 69.2%, and 68.0%, respectively, where the AUC value is 2.1% higher than the baseline algorithm TrDiMP (Transformer Discriminative Model Prediction). On other datasets, our tracker also achieves excellent tracking performance.

Color and Attention for U : Modified Multi Attention U-Net for a Better Image Colorization

Image colorization is a tedious task that requires creativity and understanding of the image context and semantic information. Many models have been made by harnessing various deep learning architectures to learn the plausible colorization. With the rapid discovery of new architecture and image generation techniques, more powerful options can be explored and improved for image colorization tasks. This research explores a new architecture to colorize an image by using pre-trained embeddings on U-Net combined with several attention modules across the model. Using embeddings from a pre-trained classifier provides a high-level feature extraction from the image. Conversely, multi-attention gives a little taste of image segmentation so that the model can distinguish objects in the image and further enhance the additional information given by the pre-trained embeddings. Adversarial training is also utilized as a normalization to make the generated image more realistic. This research preferred Parch GAN over base GAN as the discriminator model to ensure that the colorization has a consistent quality across all patches. The study shows that this U-Net modification can improve the generated image quality compared to a normal U-Net. The proposed architecture reaches an FID of 48.6253, SSIM of 0.8568, and PSNR of 19.7831 by only training it for 25 epochs; hence, this research offers another view of image colorization by using modules that are often used for image segmentation tasks.

Generative language models on nucleotide sequences of human genes.

Language models, especially transformer-based ones, have achieved colossal success in natural language processing. To be precise, studies like BERT for natural language understanding and works like GPT-3 for natural language generation are very important. If we consider DNA sequences as a text written with an alphabet of four letters representing the nucleotides, they are similar in structure to natural languages. This similarity has led to the development of discriminative language models such as DNABERT in the field of DNA-related bioinformatics. To our knowledge, however, the generative side of the coin is still largely unexplored. Therefore, we have focused on the development of an autoregressive generative language model such as GPT-3 for DNA sequences. Since working with whole DNA sequences is challenging without extensive computational resources, we decided to conduct our study on a smaller scale and focus on nucleotide sequences of human genes, i.e. unique parts of DNA with specific functions, rather than the whole DNA. This decision has not significantly changed the structure of the problem, as both DNA and genes can be considered as 1D sequences consisting of four different nucleotides without losing much information and without oversimplification. First of all, we systematically studied an almost entirely unexplored problem and observed that recurrent neural networks (RNNs) perform best, while simple techniques such as N-grams are also promising. Another beneficial point was learning how to work with generative models on languages we do not understand, unlike natural languages. The importance of using real-world tasks beyond classical metrics such as perplexity was noted. In addition, we examined whether the data-hungry nature of these models can be altered by selecting a language with minimal vocabulary size, four due to four different types of nucleotides. The reason for reviewing this was that choosing such a language might make the problem easier. However, in this study, we found that this did not change the amount of data required very much.

Detection of Chicken Freshness Utilizing VNIR, SWIR Spectroscopy, and Data Fusion

In light of the escalating demand for enhanced chicken quality and safety, there is an imperative need for an advanced methodology that can accurately and expeditiously ascertain the freshness of chicken. This study endeavors to harness hyperspectral imaging (HSI) technology, in synergy with machine learning and deep learning algorithms, to innovate a non-destructive method for the assessment of chicken freshness. In this study, chicken freshness was categorized into three distinct levels based on a comprehensive range of evaluation criteria specific to chicken freshness. Subsequent to preprocessing the spectra data, a discriminative model for chicken freshness predicated on Visible and Near-Infrared (VNIR,400-1000 nm) and Short-Wave Infrared (SWIR, 900-1700 nm) spectra was formulated utilizing both the raw and the preprocessed datasets. Consequently, key wavelengths were discerned via feature wavelengths selecting within the full spectra wavelengths, culminating in the establishment of the feature-wavelength model. The outcomes indicate that the VNIR-ResNet model, incorporating Normalization preprocessing, outperforms other full-spectra models, boasting an accuracy rate of 98.31%. Following feature wavelengths modeling, the precision of the feature selecting augmented by CARS and SPA was enhanced to 98.87%, respectively. Subsequently, a fusion model is developed through the application of a data fusion technique, the accuracy of data-layer fusion modeling was 98.87%, surpassing that derived from a singular data source, albeit the efficacy of feature-layer fusion modeling fell short of ideal. In summary, considering factors such as the cost and volume of hyperspectral data, the models such as MN-ResNet based on VNIR and MN-CARS-ResNet based on feature selecting emerge as more cost-effective and pragmatic solutions.

Is Short-Read 16S rRNA Sequencing of Oral Microbiome Sampling a Suitable Diagnostic Tool for Head and Neck Cancer?

The oral microbiome, studied by sampling the saliva or by oral rinse, has been long thought to have diagnostic capacity for head and neck cancers (HNC). However, previous reports on the HNC oral microbiome provide inconsistent results. The aim of this study is to consolidate these datasets and determine the oral microbial composition between HNC patients to healthy and premalignant individuals. We analyzed 16 published head and neck cancer (HNC) short-read 16S rRNA sequencing datasets, specifically targeting the V3V4, V4 and V4V5 regions. These datasets included saliva and oral rinse samples from donors with HNC, as well as from healthy and premalignant donors. Differences in diversities and microbial abundance were determined. HNC saliva displayed lower alpha diversity than healthy donors. In contrast, the opposite trend was observed for oral rinse samples. Beta diversity scores were largely similar across different patient types. Similar oral phyla were detected for all samples, but proportions were largely dependent on sample type (i.e., saliva or oral rinse) and primer set utilized for 16S rRNA sequencing. Neisseria, Leptotrichia and Megasphaera were elevated in healthy saliva, while Mycoplasma was elevated in HNC saliva. Oral rinse and saliva displayed similar enrichment for Fusobacterium, while Veillonella, Alloprevotella, and Campylobacter showed conflicting results. The sparse partial least squares discriminant analysis model performed effectively in discriminating HNC from healthy or premalignant patients using V3V4 saliva (AUC = 0.888) and V3V4 oral rinse (AUC = 0.928), while poor discriminative capacity was observed for V4 saliva (AUC = 0.688). In conclusion, our meta-analysis highlighted the limitations of 16S rRNA sequencing, particularly due to variations across study batches, primer sets (i.e., V3V4, V4), and sample types. Hence, caution should be exercised when interpreting 16S rRNA sequencing results across studies, especially when different primer sets and sample types are used.