Improved cytokine\u2013receptor interaction prediction by exploiting the negative sample space

Contrastive learning is an important technique in the field of machine learning for learning data representations. In the field of self-supervised visual representation learning, the strategy of positive and negative sample selection is key to improving the efficiency and effectiveness of model learning. Traditional self-supervised learning methods often employ random sampling strategies to select positive and negative samples, but this approach may result in uneven sample quality when dealing with complex datasets, thereby affecting the learning outcomes. To alleviate this problem, this study is dedicated to exploring more effective strategies for positive and negative sample selection and processing to optimize the self-supervised learning process. To this end, we propose an improved method of self-supervised learning called contrastive learning with enhanced diversity. On the one hand, this method utilizes the weight parameters of the DINO pre-trained model to initialize the feature extraction network of SimCLR, providing more accurate calculations of feature similarity. On the other hand, by setting a threshold on the feature similarity matrix and penalizing (subtracting 0.5 from) similarity scores that do not exceed this threshold, we reduce the excessive impact of high similarity scores on model training, thereby helping the model better distinguish between positive and negative samples. In downstream image classification tasks, we conducted detailed evaluations of the improved model, specifically including fine-tuning and linear evaluation aspects. Experiments show that the proposed approach improves the performance of the loss functions and improves the accuracy of the proposed SimCLR model.

In recent years, there have been rapid improvements in both remote sensing methods and satellite image availability that have the potential to massively improve burn severity assessments of the Alaskan boreal forest. In this study, we utilized recent pre- and post-fire Sentinel-2 satellite imagery of the 2019 Nugget Creek and Shovel Creek burn scars located in Interior Alaska to both assess burn severity across the burn scars and test the effectiveness of several remote sensing methods for generating accurate map products: Normalized Difference Vegetation Index (NDVI), Normalized Burn Ratio (NBR), and Random Forest (RF) and Support Vector Machine (SVM) supervised classification. We used 52 Composite Burn Index (CBI) plots from the Shovel Creek burn scar and 28 from the Nugget Creek burn scar for training classifiers and product validation. For the Shovel Creek burn scar, the RF and SVM machine learning (ML) classification methods outperformed the traditional spectral indices that use linear regression to separate burn severity classes (RF and SVM accuracy, 83.33%, versus NBR accuracy, 73.08%). However, for the Nugget Creek burn scar, the NDVI product (accuracy: 96%) outperformed the other indices and ML classifiers. In this study, we demonstrated that when sufficient ground truth data is available, the ML classifiers can be very effective for reliable mapping of burn severity in the Alaskan boreal forest. Since the performance of ML classifiers are dependent on the quantity of ground truth data, when sufficient ground truth data is available, the ML classification methods would be better at assessing burn severity, whereas with limited ground truth data the traditional spectral indices would be better suited. We also looked at the relationship between burn severity, fuel type, and topography (aspect and slope) and found that the relationship is site-dependent.

Active learning is one form of supervised machine learning. In supervised learning, a set of labeled samples is passed to a learning algorithm for training a classifier. However, labeling large amounts of training samples can be costly and error-prone. Active learning deals with the development of algorithms that interactively select a subset of the available unlabeled samples for labeling, and aims at minimizing the labeling effort while maintaining classification performance. The key challenge for the development of so-called active learning strategies is the balance between exploitation and exploration: On the one hand, the estimated decision boundary needs to be refined in feature space regions where it has already been established, while, on the other hand, the feature space needs to be scanned carefully for unexpected class distributions. In this thesis, two approaches to active learning are presented that consider these two aspects in a novel way. In order to lay the foundations for the first one, it is proposed to express the uncertainty in class prediction of a classifier at a test point in terms of a second-order distribution. The mean of this distribution corresponds to the common estimate of the posterior class probabilities and thus is related to the distance of the test point to the decision boundary, whereas the spread of the distribution indicates the degree of exploration in the corresponding region of feature space. This allows for the evaluation of the utility of labeling a yet unlabeled point with respect to classifier improvement in a principled way and leads to a completely novel approach to active learning. The proposed strategy is then implemented and evaluated based on kernel density classification. The generic active learning strategy can be combined with any other classifier, but it performs best if the derived second-order distributions are sufficiently good approximations to the sampling distribution. Although second-order distributions for random forests are derived in this thesis, they do not approximate sufficiently well the sampling distribution and mainly allow only for the relative comparison of prediction uncertainty between test points. In order to combine the state of the art classification performance of random forests with the principal ideas of the first active learning approach, a related second approach for random forests is derived. It is, in addition, tailored to the demands in industrial optical inspection: bag-wise labeling with weak labels and strongly imbalanced classes. Moreover, an outlier detection scheme based on random forests is derived that is used by the proposed active learning algorithm. Finally, a new computational scheme for Gaussian process classification is presented. It is compared to two standard methods in geostatistics, both with respect to theoretical consistency and practical performance. The method evolved as a by-product when considering using Gaussian process models for active learning.

Jiangxi Province, characterized by abundant forest resources and complex topography, is highly susceptible to forest fires. This study integrated multiple factors, including topography, climate, vegetation, and human activities, and employed machine learning models, specifically random forest (RF), support vector machine (SVM), and back-propagation neural network (BPNN), to predict forest fire occurrence in Jiangxi. Using Moderate Resolution Imaging Spectroradiometer L3 fire-point data from 2001–2020, we analyzed the spatiotemporal distribution of forest fires and applied the weight of evidence (WoE) method to evaluate the correlation between forest fires and environmental factors. WoE was employed to select negative samples, which were compared with those obtained using traditional random sampling methods. The optimal model was then utilized to generate seasonal spatial distribution maps of forest fire risk throughout Jiangxi Province. The results showed that over the past two decades, the frequency of forest fires generally decreased. RF demonstrated a significant advantage over SVM and BPNN in predicting forest fires. Vegetation coverage was the most influential factor. In addition, the models trained with WoE-selected negative samples exhibited enhanced accuracy, with area under the curve values increasing from 0.946 to 0.995 for RF, 0.8344 to 0.925 for SVM, and 0.832 to 0.850 for BPNN, compared to those trained with randomly sampled negative data. Finally, forest fires were most frequent during winter, particularly in Ganzhou, Fuzhou, and Ji'an. High-risk fire zones were more dispersed in spring, whereas autumn fires were primarily concentrated in Ganzhou, and fire activity was relatively low during summer. The seasonal forest fire risk maps generated in this study offer valuable insights for guiding forest fire management in the Jiangxi Province and similar regions, providing critical practical significance for informed decision-making.

This study investigates the performance of machine learning classifiers in the domain of speaker identification, a pivotal component of modern digital security systems. With the burgeoning integration of voice-activated interfaces in technology, the demand for accurate and reliable speaker identification is paramount. This research provides a comprehensive comparison of four widely used classifiers: Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbors (KNN), and Decision Tree (DT). Utilizing the LibriSpeech dataset, known for its diversity of speakers and recording conditions, we extracted Mel-frequency cepstral coefficients (MFCCs) to serve as features for training and evaluating the classifiers. Each model's performance was assessed based on precision, recall, F1-score, and accuracy. The results revealed that RF outperformed all other classifiers, achieving near-perfect metrics, indicative of its robustness and generalizability for speaker identification tasks. KNN also demonstrated high performance, suggesting its suitability for applications where rapid execution and interpretability are critical. Conversely, SVM and DT, while yielding moderate and lower performances respectively, highlighted the necessity for further optimization. These findings underscore the effectiveness of ensemble and distance-based classifiers in handling complex patterns for speaker differentiation. The study not only guides the selection of appropriate classifiers for speaker identification but also sets the stage for future research, which could explore hybrid models and the impact of dataset variability on performance. The insights from this analysis contribute significantly to the field, providing a benchmark for developing advanced speaker identification systems

Few-shot relation extraction aims to identify and classify specific semantic relations between entities from text with a small number of annotated examples. Recent studies have shown that innovative model designs and learning strategies can significantly enhance the model’s generalization ability and performance, even with limited training samples. However, the samples used for training face the dual challenges of scarce annotated data and semantic ambiguity. Given the limitations of the existing SaCon framework in negative sample generation and discrimination efficiency, we propose a dynamic adversarial negative sample enhancement strategy. This strategy introduces adversarial perturbations before encoding the pre-trained language model by constructing a multi-granularity semantic space. Specifically, we first design an entity permutation mechanism to randomly exchange the subject/object entities of sentences in a small batch to generate negative sample clusters with similar semantics but misplaced relations, then, we integrate a multi-view contrastive learning framework to embed adversarial samples into the feature space topology optimization process.To strengthen boundary-sensitive features, an adaptive margin ranking loss function is proposed to dynamically adjust the representation distance constraints of positive and negative samples, forcing the model to capture the deep semantic invariance of relational predicates under limited samples. This method aims to optimize the traditional negative sample random sampling paradigm, actively explore the semantic space through an adversarial generation mechanism, and construct “difficult samples” with minimal semantic deviation through gradient back propagation, thereby improving the model’s ability to parse implicit relational patterns. The experiment result shows that this framework effectively alleviates the risk of overfitting in small sample scenarios by decoupling relational semantics and surface syntactic features, and its dynamic loss design provides mathematical guarantees for orthogonal separation of feature space. Our code is available online at https://github.com/hhy-test/ESCR.

Landslide susceptibility assessment (LSA) plays a crucial role in disaster prevention and mitigation. Traditional random selection of non-landslide samples (labeled as 0) suffers from poor representativeness and high randomness, which may include potential landslide areas and affect the accuracy of LSA. To address this issue, this study proposes a novel Landslide Susceptibility Index–based Semi-supervised Fuzzy C-Means (LSI-SFCM) sampling strategy combining membership degrees. It utilizes landslide and unlabeled samples to map landslide membership degree via Semi-supervised Fuzzy C-Means (SFCM). Non-landslide samples are selected from low-membership regions and assigned membership values as labels. This study developed three models for LSA—Convolutional Neural Network (CNN), U-Net, and Support Vector Machine (SVM), and compared three negative sample sampling strategies: Random Sampling (RS), SFCM (samples labeled 0), and LSI-SFCM. The results demonstrate that the LSI-SFCM effectively enhances the representativeness and diversity of negative samples, improving the predictive performance and classification reliability. Deep learning models using LSI-SFCM performed with superior predictive capability. The CNN model achieved an area under the receiver operating characteristic curve (AUC) of 95.52% and a prediction rate curve value of 0.859. Furthermore, compared with the traditional unsupervised fuzzy C-means (FCM) clustering, SFCM produced a more reasonable distribution of landslide membership degrees, better reflecting the distinction between landslides and non-landslides. This approach enhances the reliability of LSA and provides a scientific basis for disaster prevention and mitigation authorities.

Human activity recognition (HAR) is a popular field of study. The outcomes of the projects in this area have the potential to impact on the quality of life of people with conditions such as dementia. HAR is focused primarily on applying machine learning classifiers on data from low level sensors such as accelerometers. The performance of these classifiers can be improved through an adequate training process. In order to improve the training process, multivariate outlier detection was used in order to improve the quality of data in the training set and, subsequently, performance of the classifier. The impact of the technique was evaluated with KNN and random forest (RF) classifiers. In the case of KNN, the performance of the classifier was improved from 55.9% to 63.59%.

This paper compares the classification performance of machine learning classifiers vs. deep learning-based handcrafted models and various pretrained deep networks. The proposed study performs a comprehensive analysis of object classification techniques implemented on low-altitude UAV datasets using various machine and deep learning models. Multiple UAV object classification is performed through widely deployed machine learning-based classifiers such as K nearest neighbor, decision trees, naïve Bayes, random forest, a deep handcrafted model based on convolutional layers, and pretrained deep models. The best result obtained using random forest classifiers on the UAV dataset is 90%. The handcrafted deep model's accuracy score suggests the efficacy of deep models over machine learning-based classifiers in low-altitude aerial images. This model attains 92.48% accuracy, which is a significant improvement over machine learning-based classifiers. Thereafter, we analyze several pretrained deep learning models, such as VGG-D, InceptionV3, DenseNet, Inception-ResNetV4, and Xception. The experimental assessment demonstrates nearly 100% accuracy values using pretrained VGG16- and VGG19-based deep networks. This paper provides a compilation of machine learning-based classifiers and pretrained deep learning models and a comprehensive classification report for the respective performance measures.

Heart disease is a leading cause of death worldwide, and early detection is crucial for improving patient outcomes. Machine learning classifiers have shown promise in predicting heart disease using patient-specific factors such as demographic information, medical history, and lifestyle habits. In this paper, we aimed to evaluate the performance of machine learning classifiers in predicting heart disease using a combination of 5 different datasets such as Cleveland, Hungarian, Switzerland, Long Beach VA, and StatLog (Heart) Datasets available on IEEE data port. The two significant challenges are addressed in this work: 1) predicting heart failure using machine-learning models without eliminating any clinical features, which increases the risk of overfitting and can result in poor performance metrics, and 2) we propose a model that will provide remarkable accuracy regardless of the type of data, to offer model generalizability. Machine-learning algorithms such as logistic regression, decision trees, random forests, support vector machine, extreme gradient boosting, extra-tree, and K-nearest neighbor are applied for heart disease prediction. In addition, ensemble approaches such as majority voting, boosting, bagging, and stacking are employed. The performance of the classifiers is evaluated using performance metrics such as accuracy, precision, recall, and F1-score. The results showed that the ensemble approach of stacking outperformed individual models, with an accuracy of 93.67%.

Reliable TCR-epitope binding prediction models are essential for development of adoptive T cell therapy and vaccine design. These models often struggle with false positives, which can be attributed to the limited data coverage in existing negative sample datasets. Common strategies for generating negative samples, such as pairing with background T cell receptors (TCRs) or shuffling within pairs, fail to account for model-specific vulnerabilities or biologically implausible sequences. To address these challenges, we propose an iterative attack-and-defend framework that systematically identifies and mitigates weaknesses in TCR-epitope prediction models. During the attack phase, a reinforcement learning from AI feedback (RLAIF) framework is used to attack a prediction model by generating biologically implausible sequences that can easily deceive the model. During the defense phase, these identified false positives are incorporated into fine-tuning dataset, enhancing the model's ability to detect false positives. A comprehensive adversarial negative dataset can be obtained by iteratively attacking and defending the model. This dataset can be directly used to improve model robustness, eliminating the need for users to conduct their own attack-and-defend cycles. We apply our framework to five existing binding prediction models, spanning diverse architectures and embedding strategies to show its efficacy. Experimental results show that our approach significantly improves these models' ability to detect adversarial false positives. The combined dataset constructed from these experiments also provides a benchmarking tool to evaluate and refine prediction models. Our framework offers a new approach for improving model robustness in other biological tasks where negative sampling is inherently limited. The curated dataset and code are available at a public repository (https://github.com/Lee-CBG/BAP_Attack_n_Defend).

Sparse sampled visualization of ultrasonic guided waves for defect identification in plate structures

Magnetic resonance imaging (MRI) scan analysis is an effective tool that accurately detects abnormal brain tissue. This manuscript proposes the strategy of segmentation of brain tumors in MRI images and uses the technique of weighted fuzzy factor based on kernel metrics. Here, a deep auto encoder (DAE) with barnacle mating algorithm (BMOA) and random forest (RF) classifier are used to tumor stage classification to enhance the accuracy of prediction. This manuscript presents a deep‐neural network structure, integrating DAE and RF, with a classification unit, which is used for the classification of brain MRI. Finally, the segmented features are graded by the DAE with BMOA and RF. The proposed method is executed in MATLAB site and the performance is analyzed with existing methods. The experimental outcomes of the proposed method are assessed and validated in MR brain images depending on accuracy, sensitivity, and specificity for performance with quality analysis.

Knowledge representation learning can realize the representation of entity and relation semantic information by projecting the triple entity or relation in knowledge graph into vector space, and can effectively calculate entity, relation and their complex semantic association. However, most of the existing models only use the structure information of triples, ignoring the semantic rich entity description. In addition, knowledge representation learning methods usually use negative sampling method. Through random sampling, the negative samples generated by this method are easy to be recognized by the model, and with the iteration of training, the contribution to the improvement of the model performance becomes smaller and smaller. Therefore, we propose a Knowledge Representation learning method based on Generative adversarial network and Pre-training language model (KRGP). Specifically, it uses the idea of generative adversarial network, semantic matching model and translation distance model to generate high-quality negative samples. We combine descriptions of head and tail entities and relations corresponding to negative sample triples into text sequences, and then use next sentence prediciton task of BERT to transform text sequences into a binary classification problem. Experimental results show that this method can improve the performance of link prediction and triple classification on multiple data sets.KeywordsKnowledge representation learningGenerative adversarial networkEntity descriptionPre-training model

A number of computational methods have been proposed that predict protein-protein interactions (PPIs) based on protein sequence features. Since the number of potential non-interacting protein pairs (negative PPIs) is very high both in absolute terms and in comparison to that of interacting protein pairs (positive PPIs), computational prediction methods rely upon subsets of negative PPIs for training and validation. Hence, the need arises for subset sampling for negative PPIs. We clarify that there are two fundamentally different types of subset sampling for negative PPIs. One is subset sampling for cross-validated testing, where one desires unbiased subsets so that predictive performance estimated with them can be safely assumed to generalize to the population level. The other is subset sampling for training, where one desires the subsets that best train predictive algorithms, even if these subsets are biased. We show that confusion between these two fundamentally different types of subset sampling led one study recently published in Bioinformatics to the erroneous conclusion that predictive algorithms based on protein sequence features are hardly better than random in predicting PPIs. Rather, both protein sequence features and the 'hubbiness' of interacting proteins contribute to effective prediction of PPIs. We provide guidance for appropriate use of random versus balanced sampling. The datasets used for this study are available at http://www.marcottelab.org/PPINegativeDataSampling. yungki@mail.utexas.edu; marcotte@icmb.utexas.edu. Supplementary data are available at Bioinformatics online.