Similar Classification Research Articles

LLM‐based Approach to Automatically Establish Traceability between Requirements and MBSE

AbstractTracing requirements specification to design and implementation is an essential part of safety standards, as it allows to ensure that safety goals are met throughout the development process. Manual tracing numerous artifacts produced throughout the development process is error‐prone and takes much time. To address these problems, we proposed a tool (Bonner, M.; Zeller, M.; Schulz, G.; Beyer, D.; Olteanu, M., 2023), which allows to establish links between requirements and Model‐Based Systems Engineering (MBSE) in a semi‐automatic way. The underlying algorithms of our tool are embedding similarity computation and classification approaches based on Large Language Models (LLMs). To assess the performance of underlying algorithm we propose an evaluation, where we compare the recall, the precision, and the F2 score of different approaches applied to our datasets. The goal of our evaluation is to understand how well LLMs perform in automatically generating trace links on different datasets. Our evaluation shows that it is worth to invest time in preprocessing the data and fine‐tuning the LLMs to achieve the better recommendations for engineers, which improves the traceability process.

호주의 불법 마약 관련 법과 통제정책:한국과의 간략한 비교

This paper provides a comprehensive overview of Australian drug laws, highlighting the evolution of laws from the late 19th century to the present, and comparing them with South Korean drug policies. It begins by exploring the historical context of Australian drug laws, highlighting the shift from minimal intervention in the late 19th century, characterized by the widespread use of drugs like opium and morphine, to more stringent drug control policies in the early 20th century. This change was influenced by international developments and domestic concerns, notably racial prejudices and the rise of international drug conventions. The paper then delves into the legal frameworks governing drug offences in Australia, explaining the interplay of federal and state laws. It outlines the specific provisions and penalties, with an emphasis on the Criminal Code Act 1995(Cth) and the definitions of serious drugs. A detailed examination of drug offences under Federal law is provided, including distribution, production, import-export and possession offences, along with their respective penalties. Furthermore, the paper discusses the enforcement and prosecution of drug laws in Australia, detailing the roles of law enforcement agencies, the prosecution process, and the judiciary's role in interpreting and applying these laws. It also highlights the shift in Australian drug policy towards harm reduction and public health approaches. This section includes case studies of harm reduction initiatives in Australia. A notable aspect of the paper is the comparison between Australian and South Korean drug laws, elucidating their similarities and differences. While both countries adopt a strict stance on illicit drugs and have similar drug classification systems, they differ in their approach to personal use and possession, harm reduction focus, legal status of cannabis, and cultural perceptions of drug use. The paper concludes by emphasizing the complexity of drug legislation and policy in Australia, shaped by international conventions and domestic priorities. It notes that Australian drug laws strike a balance between prohibition, control, and public health, providing valuable insights for international audiences. The paper also recognizes the dynamic nature of Australian drug law, suggesting that its evolution could significantly influence global drug policy, especially with emerging challenges like synthetic drugs and online drug trafficking.

경계선급 지적 기능성(BIF) 초등학생의 학교생활적응에 필요한 지원요소 탐색

The purpose of this study is to explore the support elements necessary for the adaptation of border-level intellectual functional elementary school students to school life recognized by field educators. One-on-one individual interviews were conducted with 10 public elementary school teachers and 10 teachers from alternative schools operated for border-level intellectual functional students. The results of similarity classification and importance evaluation were both visualized by a multidimensional scale analysis and hierarchical cluster analysis based on the extracted statements. As a result of this study, first, support in the relational aspect should be expanded. Support for border-level intellectual functional students, which has thus far been biased towards the learning aspect, should be provided in a balanced way when regarding the relational aspect. Second, support for establishing awareness of the educational community object is necessary; this is so that the educational community can fully understand the characteristics of border-level intellectual functionality that are distinguished from other disabilities. Third, cooperative support between several educational entities with homeroom teachers and parents as the core subjects is needed. In order for border-level intellectual functional elementary school students to adapt to school life, it should be recognized that support focus needs to be tailored towards how to create an educational environment including homes and schools surrounding them, rather than how to improve their functions.

Similarity-Based Three-Way Clustering by Using Dimensionality Reduction

Three-way clustering uses core region and fringe region to describe a cluster, which divide the dataset into three parts. The division helps identify the central core and outer sparse regions of a cluster. One of the main challenges in three-way clustering is the meaningful construction of the two sets. Aimed at handling high-dimensional data and improving the stability of clustering, this paper proposes a novel three-way clustering method. The proposed method uses dimensionality reduction techniques to reduce data dimensions and eliminate noise. Based on the reduced dataset, random sampling and feature extraction are performed multiple times to introduce randomness and diversity, enhancing the algorithm’s robustness. Ensemble strategies are applied on these subsets, and the k-means algorithm is utilized to obtain multiple clustering results. Based on these results, we obtain co-association frequency between different samples and fused clustering result using the single-linkage method of hierarchical clustering. In order to describe the core region and fringe region of each cluster, the similar class of each sample is defined by co-association frequency. The lower and upper approximations of each cluster are obtained based on similar class. The samples in the lower approximation of each cluster belong to the core region of the cluster. The differences between lower and upper approximations of each cluster are defined as fringe region. Therefore, a three-way explanation of each cluster is naturally formed. By employing various UC Irvine Machine Learning Repository (UCI) datasets and comparing different clustering metrics such as Normalized Mutual Information (NMI), Adjusted Rand Index (ARI), and Accuracy (ACC), the experimental results show that the proposed strategy is effective in improving the structure of clustering results.

Conditional Information Gain Trellis

Conditional computing processes an input using only part of the neural network’s computational units. Learning to execute parts of a deep convolutional network by routing individual samples has several advantages: This can facilitate the interpretability of the model, reduce the model complexity, and reduce the computational burden during training and inference. Furthermore, if similar classes are routed to the same path, that part of the network learns to discriminate between finer differences and better classification accuracies can be attained with fewer parameters. Recently, several papers have exploited this idea to select a particular child of a node in a tree-shaped network or to skip parts of a network. In this work, we follow a Trellis-based approach for generating specific execution paths in a deep convolutional neural network. We have designed routing mechanisms that use differentiable information gain-based cost functions to determine which subset of features in a convolutional layer will be executed. We call our method Conditional Information Gain Trellis (CIGT). We show that our conditional execution mechanism achieves comparable or better model performance compared to unconditional baselines, using only a fraction of the computational resources. We provide our code and model checkpoints used in the paper at: https://github.com/ufukcbicici/cigt/tree/prl/prl_scripts.

Compact machine learning model for the accurate prediction of first 24-hour survival of mechanically ventilated patients.

The field of machine learning has been evolving and applied in medical applications. We utilised a public dataset, MIMIC-III, to develop compact models that can accurately predict the outcome of mechanically ventilated patients in the first 24 h of first-time hospital admission. 67 predictive features, grouped into 6 categories, were selected for the classification and prediction task. 4 tree-based algorithms (Decision Tree, Bagging, eXtreme Gradient Boosting and Random Forest), and 5 non-tree-based algorithms (Logistic Regression, K-Nearest Neighbour, Linear Discriminant Analysis, Support Vector Machine and Naïve Bayes), were employed to predict the outcome of 18,883 mechanically ventilated patients. 5 scenarios were crafted to mirror the target population as per existing literature. S1.1 reflected an imbalanced situation, with significantly fewer mortality cases than survival ones, and both the training and test sets played similar target class distributions. S1.2 and S2.2 featured balanced classes; however, instances from the majority class were removed from the test set and/or the training set. S1.3 and S 2.3 generated additional instances of the minority class via the Synthetic Minority Over-sampling Technique. Standard evaluation metrics were used to determine the best-performing models for each scenario. With the best performers, Autofeat, an automated feature engineering library, was used to eliminate less important features per scenario. Tree-based models generally outperformed the non-tree-based ones. Moreover, XGB consistently yielded the highest AUC score (between 0.91 and 0.97), while exhibiting relatively high Sensitivity (between 0.58 and 0.88) on 4 scenarios (1.2, 2.2, 1.3, and 2.3). After reducing a significant number of predictors, the selected calibrated ML models were still able to achieve similar AUC and MCC scores across those scenarios. The calibration curves of the XGB and BG models, both prior to and post dimension reduction in Scenario 2.2, showed better alignment to the perfect calibration line than curves produced from other algorithms. This study demonstrated that dimension-reduced models can perform well and are able to retain the important features for the classification tasks. Deploying a compact machine learning model into production helps reduce costs in terms of computational resources and monitoring changes in input data over time.

Adaptive attribute distribution similarity for few-shot learning

Modern deep learning has many drawbacks, including a heavy reliance on labeled data. One of the key strategies for solving this problem is few-shot learning (FSL). With just a few labeled samples, FSL seeks to identify previously unknown classes. The majority of works that have been published thus far focus on comparing the features of query samples and support classes, which do not fully utilize the training set's data and do not help sustain performance improvement. In our study, we compute a new attribute distribution similarity between support classes and a query sample of novel classes using attribute information on the training set. We suggest a fresh approach to three phases to accomplish our objective: 1) A attribute provider harnesses the visual features of the training set to construct attributes. 2) Choosing appropriate attributes for novel classes and enriching attributes to determine how similar novel classes and attributes are to one another. 3) To help with classification, attribute distribution similarity is computed for the first time by creating new correlations between the support classes and the query samples, which increases the accuracy of picture classification. Be aware that our solution won't make the initial network settings larger. Experiments on inductive FSL tasks demonstrate the usefulness and practicality of our strategy. Specifically, Our method has achieved the highest performance in the 5-way 1-shot task settings on the tiered-ImageNet and CUB 200–2011 datasets, with impressive results of 73.22% and 82.34% respectively.

Description of “Ovino Belmontese”, a new semisoft sheep's milk cheese processed using “Italico” cheese technology

The objective of this study was to create a new semisoft sheep's milk cheese called “Ovino Belmontese” cheese (OBCh) by applying the “Italico” cheese-making technology. The cheese production took place under industrial conditions, with the addition of a commercial starter formulation containing Streptococcus thermophilus. The microbiological, physicochemical, and sensory characteristics of OBCh were assessed and compared to those of a commercially available cow's Italico cheese (CICh). Streptococcus thermophilus dominated the microbial community during the cheese-making process, reaching levels of approximately 9.0 Log CFU/g in both OBCh and CICh. Among physical characteristics, no statistically significant difference (p ≥ 0.05) was registered in terms of lightness, redness, yellowness, and hardness between the two cheeses. OBCh exhibited a twofold higher short-chain fatty acid content compared to CICh. Both cheeses displayed similar classes of volatile organic compounds, although their relative percentages differed. The application of Italico cheese technology to process sheep's milk did not negatively affect sensory attributes. This study highlighted that utilizing a cheese-making technology not commonly used for processing sheep's milk represents a promising strategy to diversify Sicilian dairy productions.

Unlocking the Potential: Biomarkers of Response to Antibody-Drug Conjugates.

Antibody-drug conjugates (ADCs) have reshaped the cancer treatment landscape across a variety of different tumor types. ADCs' peculiar pharmacologic design combines the cytotoxic properties of chemotherapeutic agents with the selectivity of targeted therapies. At present, the approval of many ADCs used in clinical practice has not always been biomarker-driven. Indeed, predicting ADCs' activity and toxicity through the demonstration of specific biomarkers is still a great unmet need, and the identification of patients who can derive significant benefit from treatment with ADCs may often be uncertain. With the lack of robust predictive biomarkers to anticipate primary, intrinsic resistance to ADCs and no consolidated biomarkers to aid in the early identification of treatment resistance (ie, acquired resistance), the determination of precise biologic mechanisms of ADC activity and safety becomes priority in the quest for better patient-centric outcomes. Of great relevance, whether the target antigen expression is a determinant of ADCs' primary activity is still to be clarified, and available data remain quite controversial. Antigen expression assessment is typically performed on tissue biopsy, hence only providing information on a specific tumor site, therefore unable to capture heterogeneous patterns of tumor antigen expression. Quantifying the expression of the target antigen across all tumor sites would help better understand tumor heterogeneity, whereas molecularly characterizing tumor-intrinsic features over time might provide information on resistance mechanisms. In addition, toxicity can represent a critical concern, since most ADCs have a safety profile that resembles that of chemotherapies, with often unique adverse events requiring special management, possibly because of the differential in pharmacokinetics between the small-molecule agent versus payload of a similar class (eg, deruxtecan conjugate-related interstitial lung disease). As such, the identification of robust predictive biomarkers of safety and activity of ADCs has the potential to improve patient selection and enrich the population of patients most likely to derive a substantial clinical benefit, especially in those disease settings where different ADCs happen to be approved in competing clinical indications, with undefined biomarkers to make precise decision making and unclear data on how to sequence ADCs. At this point, the identification of clinically actionable biomarkers in the space of ADCs remains a top research priority.

Integration of UML Class Diagrams Based on Semantics and Structure

As a high-level reuse, software design reuse has received more and more attention because of its important impact on the subsequent development stages. Usually, the design models are typically represented as some graphs or diagrams, in which Unified Modeling Language (UML) class diagram is so widely used in software design that it has become the de facto standard. There has been some research on the reuse of UML class diagrams so far, mainly focusing on matching and retrieval. However, it is worth noting that there are many similar class diagrams modeling the same object or some related class diagrams modeling different aspects of the same object in the reuse repository. As a matter of fact, the primary step to achieve a high-quality reuse is to have high-quality software artifacts, so the well-designed UML class diagrams become a necessary resource for software design reuse. Therefore, it is necessary to integrate these class diagrams so that they have stronger modeling ability, and eliminating redundancy is another benefit of the integration. Up to now, there has been little discussion about the integration of class diagrams, so we propose an integration approach based on semantics and structure in this paper. The equivalent elements that can identify the semantically merged parts are defined, and the possible conflict items are listed from both semantic and structural aspects. The integration procedure composed of three stages is proposed, in which an approach combining semantic common class diagrams (SCCDs) and structural common graph sequence (SCGS) is combined to determine the merged parts, the integration issue of heterogeneous class diagrams is considered from the proposed abstract models, and the conflict resolution for each conflict item is described by examples. The experimental results show the effectiveness of our proposed integration approach.

Black Americans' Diminished Health Returns of Professional Occupations: A Thirty-Year Follow-Up Study of Middle-Aged and Older Adults.

Occupational classes exert substantial effects on both subjective and objective health outcomes. However, it remains unclear whether the health impact of similar occupational classes varies across racial groups among middle-aged and older adults in the United States. Grounded in the theory of Minorities' Diminished Returns (MDRs), which posits that health benefits from resources such as employment are systematically weaker for racial minority populations, particularly Non-Latino Black individuals, this study tested Black-White disparities in the effects of similar occupational classes on health outcomes in middle-aged and older adults. Utilizing data from the Health and Retirement Study (HRS), we employed a 30-year longitudinal design with a nationally representative sample of middle-aged and older adults in the United States. Six occupational classes-operator, managerial, professional specialty, sales, clerical/admin, and service-served as the key predictor variables (independent variables), with race as the moderator. Various health outcomes, including self-rated health, chronic disease, body mass index (BMI), activities of daily living (ADL), and cognitive function, were measured longitudinally from wave 1 to wave 15 (from baseline to 30 years later). Statistical analyses, incorporating logistic regression models, were conducted to assess associations between occupational class and health outcomes overall and based on race. Our analysis included 7538 Non-Latino White or Non-Latino Black participants followed for up to 30 years. Initial findings revealed positive health effects of professional occupations on cognitive function and self-rated health over 30 years of follow-uWe also identified significant interactions between race and professional occupational class on all health outcomes, indicating notable racial differences in the effects of professional occupations on health outcomes across domains. The effects of professional occupational class were weaker for Non-Latino Black individuals than for Non-Latino White individuals. Consistent with the Minorities' Diminished Returns theory, our findings indicated less pronounced positive effects of professional occupations on a wide range of health outcomes for Non-Latino Black individuals compared to Non-Latino Whites. These disparities emphasize the critical need to address structural factors that contribute to the diminished returns of prestigious occupations for Non-Latino Black populations.

Combining “Deep Learning” and Physically Constrained Neural Networks to Derive Complex Glaciological Change Processes from Modern High-Resolution Satellite Imagery: Application of the GEOCLASS-Image System to Create VarioCNN for Glacier Surges

The objectives of this paper are to investigate the trade-offs between a physically constrained neural network and a deep, convolutional neural network and to design a combined ML approach (“VarioCNN”). Our solution is provided in the framework of a cyberinfrastructure that includes a +newly designed ML software, GEOCLASS-image (v1.0), modern high-resolution satellite image data sets (Maxar WorldView data), and instructions/descriptions that may facilitate solving similar spatial classification problems. Combining the advantages of the physically-driven connectionist-geostatistical classification method with those of an efficient CNN, VarioCNN provides a means for rapid and efficient extraction of complex geophysical information from submeter resolution satellite imagery. A retraining loop overcomes the difficulties of creating a labeled training data set. Computational analyses and developments are centered on a specific, but generalizable, geophysical problem: The classification of crevasse types that form during the surge of a glacier system. A surge is a glacial catastrophe, an acceleration of a glacier to typically 100–200 times its normal velocity. GEOCLASS-image is applied to study the current (2016-2024) surge in the Negribreen Glacier System, Svalbard. The geophysical result is a description of the structural evolution and expansion of the surge, based on crevasse types that capture ice deformation in six simplified classes.

Drug-Leachable Interaction Product Evaluation in Prefilled Syringe of Ganirelix Acetate Injection

The concept of extractables and leachable has introduced a new era for identifying potential impurities in drug products. Pharmaceutical packaging materials encompass a variety of polymers due to their appealing properties for storing the drug product. However, numerous chemical species may leach into the drug product from these polymers, posing significant health hazards in the public domain. Identifying such leachable is crucial for assessing safety and addressing toxicological concerns. Acrylic acids are commonly used materials for adhering needles to the barrel in pre-filled syringes. In this study, we identified acrylic acid leachable impurities in ganirelix drug products available in the market using mass spectrometry as an analytical technique. These impurities leached into the drug product during storage conditions. Characterization of the impurities was carried out through data obtained from HRMS and MS/MS analysis, revealing them as polyacrylic acid-ganirelix adduct impurities. This study offers valuable insights into identifying leachable and susceptible sites, providing a foundation for potential modifications in similar classes of drug products, thereby enhancing their safety and efficacy.

Warning signs in a period of acute risk for suicide attempt: The utility of count- and combination-based classification

ObjectiveThis study sought to establish if models involving the specific combinations of identifiable behavioral warning signs (WS; i.e., alcohol use, suicidal communications, preparation of personal affairs, experiencing negative interpersonal life events) for a suicide attempt outperformed a simpler count model of WS to distinguish an acute risk period (the six hours prior to a suicide attempt) from a control period (a matched six-hour period the day prior). MethodGeneralized linear mixed models tested all logical combinations of four behavioral WS in addition to a count variable of the number of behavioral warning signs. For the count variable, receiver operating characteristic curve analysis was utilized to determine the optimal cut point for the number of endorsed WS. ResultsArea under the curve (AUC) compared classification across all models, with the count model demonstrating similar classification performance to the best multivariate model (AUCs of 0.72 and 0.73, respectively), with an optimal cut point of endorsing one or more WS. ConclusionAlthough determinations of acute risk should be informed by multiple sources of information, this study suggests a relatively simple count-based approach that considers the presence of one (or more) behavioral WS may be used as a potential indication of increased acute risk for suicide attempt.

Boosting few-shot rare skin disease classification via self-supervision and distribution calibration.

Due to the difficulty in obtaining clinical samples and the high cost of labeling, rare skin diseases are characterized by data scarcity, making training deep neural networks for classification challenging. In recent years, few-shot learning has emerged as a promising solution, enabling models to recognize unseen disease classes by limited labeled samples. However, most existing methods ignored the fine-grained nature of rare skin diseases, resulting in poor performance when generalizing to highly similar classes. Moreover, the distributions learned from limited labeled data are biased, severely impairing the model's generalizability. This paper proposes a self-supervision distribution calibration network (SS-DCN) to address the above issues. Specifically, SS-DCN adopts a multi-task learning framework during pre-training. By introducing self-supervised tasks to aid in supervised learning, the model can learn more discriminative and transferable visual representations. Furthermore, SS-DCN applied an enhanced distribution calibration (EDC) strategy, which utilizes the statistics of base classes with sufficient samples to calibrate the bias distribution of novel classes with few-shot samples. By generating more samples from the calibrated distribution, EDC can provide sufficient supervision for subsequent classifier training. The proposed method is evaluated on three public skin disease datasets(i.e., ISIC2018, Derm7pt, and SD198), achieving significant performance improvements over state-of-the-art methods.

Pyrazolidinone-Aided Ru(II)-Catalyzed Regioselective C-H Annulation with Allenes.

Regioselective annulation of allenes via C-H activation represents an elegant synthetic approach toward the construction of valuable scaffolds. Considering the importance of allenes, herein we developed an unprecedented Ru(II)-catalyzed highly regioselective redox-neutral C-H activation/(4 + 1)-annulation of 1-arylpyrazolidinones employing allenyl acetates to access pyrazolo[1,2-a]indazol-1-one derivatives. Additionally, allenyl cyclic carbonates, which were never tested in C-H activation, were utilized to construct a similar class of heterocycles having a pendent alcohol functionality. Notably, double C-H functionalization was achieved by a simple modification of reaction conditions. The synthetic significance of this methodology is underscored by late-stage modification of natural products, broad substrate scope, gram-scale synthesis, and postfunctionalizations.

Nanoconfinement Effect in Water Processable Discrete Molecular Complex-Based Hybrid Piezo- and Thermo-Electric Nanogenerator.

Water-processable hybrid piezo- and thermo-electric materials have an increasing range of applications. We use the nanoconfinement effect of ferroelectric discrete molecular complex [Cu(l-phe)(bpy)(H2O)]PF6·H2O (1) in a nonpolar polymer 1D-nanofiber to envision the high-performance flexible hybrid piezo- and thermo-electric nanogenerator (TEG). The 1D-nanoconfined crystallization of 1 enhances piezoelectric throughput with a high degree of mechano-sensitivity, i.e., 710 mV/N up to 3 N of applied force with 10,000 cycles of unaffected mechanical endurance. Thermoelectric properties analysis shows a noticeable improvement in Seebeck coefficient (∼4 fold) and power factor (∼6 fold) as compared to its film counterpart, which is attributed to the enhanced density of states near the Fermi edges as evidenced by ultraviolet photoelectric spectroscopy and density functional based theoretical calculations. We report an aqueous processable hybrid TEG that provides an impressive magnitude of Seebeck coefficient (∼793 μV/K) and power factor (∼35 mWm-1K-2) in comparison to a similar class of materials.

Consistent land use and land cover classification across 20 years of various high-resolution images for detecting soil sealing in Murcia, Spain

Artificial human-induced soil sealing has numerous negative consequences. The extent of impervious surfaces is a key indicator of the location and intensity of human activity; however, it is also proof of damage to the natural environment as a result of the sealing and modification of ecosystems. Remote sensing techniques can help detect and monitor changes in land use and cover over an extended period. However, the limited availability of consistent satellite images with high spatiotemporal resolutions covering several decades poses major challenges for achieving high overall classification accuracy. An accurate methodology for the multitemporal detection of artificial land cover classes was developed and applied to a case study of the metropolitan area of Murcia (Spain) with its challenging landscape conditions due to the frequent presence of bare soil. For this purpose, a variety of high-resolution satellite images from SPOT 5, Rapid Eye, and PlanetScope covering a period of 20 years were used. To improve the automated detection of built-up areas, the reflectance values of the images, normalised difference vegetation index (NDVI) and soil adjusted vegetation index (SAVI), and a building surface digital model were used as inputs for the supervised classification model. We applied a random forest algorithm to non-public, high-resolution images in the Google Earth Engine (GEE) as a processing environment to identify eight target land cover classes. The results show that the proposed methodology leads to a substantial improvement, after including the indices and the digital building model, in the overall accuracy (from 93.16 to 95.97%) and in all classes. This improvement was significant for the artificial classes and was particularly noticeable for the built-up areas (from 91.1 to 95.64%) because their confusion with bare soil was considerably reduced. This work demonstrates the effectiveness of the building-surface digital model as a tool for training the classification model, as it reduces uncertainty in confusion with other spectrally similar classes and its applicability to multisource imagery.

Transductive zero-shot learning with generative model-driven structure alignment

Zero-shot learning (ZSL) facilitates the transfer of knowledge from seen to unseen categories through high-dimensional vectors that capture both known and unknown class names. However it encounters challenges with domain shift arising from a lack of sufficient labeled data. Although transductive zero-shot learning (TZSL) addresses this bias by including samples from unseen classes, it still faces obstacles in enhancing TZSL performance. In this study, We introduce the Structure Alignment Variational Autoencoder Generative Adversarial Network (SA-VAEGAN), a novel approach that enhances the alignment between visual and auxiliary spaces. We delved into the underlying causes of domain shift and introduced a structural alignment (SA) strategy to tackle these challenges. The SA model thoroughly accounts for both inter-class and intra-class dynamics, designed to leverage the model’s comprehension of high-level semantic relations to disambiguate confusion among similar classes and mitigate intra-class confusion by penalizing atypical visual samples within classes. Assessed across four benchmark datasets, SA-VAEGAN has established a new performance standard, underscoring its efficiency in addressing the domain shift challenge within TZSL tasks, and achieving high accuracy.

A feature selection method for multimodal multispectral LiDAR sensing

Optical remote sensing techniques can indicate the properties of objects by observing different modalities (physical quantities) of the backscattered light at different optical wavelengths. Established examples are reflectance, fluorescence, Raman, or depolarization spectroscopy. LiDAR sensing, on the other hand, allows acquiring the geometry of objects by measuring the propagation delay of optical probing signals. Multimodal multispectral (MM) LiDAR combines these capabilities and extends conventional monochromatic LiDAR in both spectral and modal dimensions within a single instrument, thus enriching point cloud data with non-geometric information. The potentially high dimension of MM LiDAR data, however, poses significant challenges for instrumental design, data acquisition, and data processing. MM LiDAR data are structured as several or all modalities are available in each of the spectral channels. The above challenges can thus be mitigated by feature selection (FS), if the structure of the features is taken into account, i.e., if entire spectral channels or entire modalities are selected or omitted. Herein, we focus on the feature selection method for MM LiDAR and propose a multiclass group feature selection algorithm (MGSVM FS) consisting of a structural sparsity-based embedded feature selection method with an all-in-one support vector machine (SVM). It tackles jointly the challenges arising from the high dimension of the MM data and the need for a multiclass classification task while exploiting the structure of the MM data. In addition, we introduce a complete workflow for evaluating the feature selection and for decision-making. We apply the framework to selecting an optimum spectral and modal configuration for remote material classification using an experimental MM LiDAR system that provides reflectance, distance, and degree of linear polarization in 28 spectral channels of 10 nm width. For the experimental investigation, we use MM LiDAR data obtained in a controlled lab environment from thirty specimens of four material classes relevant for construction. Using all three modalities, we find a configuration with only 3 spectral channels that achieves a classification mean-F1 score of 100% within this small dataset. Similar classification performance can also be achieved with only two modalities when using more spectral channels. MGSVM FS improves the classification mean-F1 score by up to 25% as compared to random selection and outperforms two other commonly used filter and embedded feature selection methods, in this application example. The proposed group feature selection algorithm and decision-making are useful for MM LiDAR, providing a link between instrumental design, data acquisition, and data processing. However, they are also transferable to other application fields related to multiclass classification, regression, and knowledge discovery, with features structured in groups. The collected MM feature dataset, the MGSVM FS algorithm, and the evaluation pipeline are accessible online.11https://github.com/yuhan-yhyh/Dataset_Code_MGSVM-FS-MM-LiDAR.git.