Internal Features Research Articles

Semantic Augmentation in Chinese Adversarial Corpus for Discourse Relation Recognition Based on Internal Semantic Elements

This paper proposes incorporating linguistic semantic information into discourse relation recognition and constructing a Semantic Augmented Chinese Discourse Corpus (SACA) comprising 9546 adversative complex sentences. In adversative complex sentences, we suggest a quadruple (P, Q, R, Qβ) representing internal semantic elements, where the semantic opposition between Q and Qβ forms the basis of the adversative relationship. P denotes the premise, and R represents the adversative reason. The overall annotation approach of this corpus follows the Penn Discourse Treebank (PDTB), except for the classification of senses. We combined insights from the Chinese Discourse Treebank (CDTB) and obtained eight sense categories for Chinese adversative complex sentences. Based on this corpus, we explore the relationship between sense classification and internal semantic elements within our newly proposed Chinese Adversative Discourse Relation Recognition (CADRR) task. Leveraging deep learning techniques, we constructed various classification models and the model that utilizes internal semantic element features, demonstrating their effectiveness and the applicability of our SACA corpus. Compared with pre-trained models, our model incorporates internal semantic element information to achieve state-of-the-art performance.

From a Piece of Paper to Court Evidence: The Means Of Collection and Examination of Physical Documents in Ukrainian Criminal Justice

The purpose of this article is to outline criminal procedural and forensic means of collection and examination of physical documents in Ukrainian criminal justice. The linguistic analysis, formal-legal, formal-logical, modeling, forecasting methods, as well as the praxeological approach in forensic science are used. The authors argue that a document, as a material object, must be obtained in a manner permitted by law and examined using the necessary forensic means in order to acquire evidentiary status and be suitable for use in criminal proceedings. It was established that the procedural status of a document depends on its internal and external features. If the content of the document carries evidentiary information, it is considered written evidence; if such information is carried by its external (material) features, it is physical (real) evidence. Procedural means of collecting documents are defined by Art. 93 of the Criminal Procedure Code of Ukraine and differ for the parties to the proceedings. Preliminary, judicial and expert examination of documents is distinguished. The main stages of the preliminary examination of documents and the necessary technical means are described. It is emphasized that Ukrainian judges avoid using their own special knowledge when examining documents. Some features of the appointment of forensic examinations of documents and evaluation of their conclusions are highlighted. The perspective of the study of methods of using documents as evidence has been established.

A retrospective study comparing the efficacy of microwave ablation and stereotactic body radiotherapy in colorectal cancer lung metastases.

The purpose of the present study was to assess and compare the efficacy of microwave ablation (MWA) and stereotactic body radiotherapy (SBRT) in the treatment of lung metastases from patients with colorectal cancer (CRC) and to identify the preferable treatment modality based on patient and tumor characteristics. Records of 118 patients with CRC with a total of 307 lung metastases who underwent SBRT or MWA between January 2015 and December 2022 were retrospectively analyzed, including the essential clinicopathological information on patients (age, sex and underlying diseases), diagnosis and treatment information [primary tumor site, levels of carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9], imaging data [diameter of lung metastasis, location of the metastasis (i.e., whether or not the tumor was adjacent to the vessel or bronchus) and internal features] and follow-up data (postoperative therapy, complications or adverse effects and survival outcomes). For statistical analysis of the local tumor progression (LTP), disease-free survival and overall survival (OS) rates, Cox regression analysis, along with the Kaplan-Meier method adjusted using inverse probability of treatment weighting (IPTW), were performed. The median follow-up duration in the present study was 31.5 months. Multivariable Cox regression analysis revealed that the CEA level, metastasis diameter and internal features were independent predictors of OS. In the IPTW-adjusted analysis, no significant difference in the 1-year OS rate was observed between the SBRT and MWA groups (92.9 vs. 93.9%; P=0.483); however, a notable discrepancy in the treatment modalities was noted, leading to significant differences in the 2- and 3-year OS rates (65.9 vs. 57.6%, P=0.001, and 44.7 vs. 36.4%, P<0.001, respectively). A significant interaction effect for the treatment modality was observed for LTP (P=0.021). In conclusion, the present study revealed that SBRT and MWA have similar therapeutic effects in terms of prolonging the survival of patients with CRC with lung metastases; however, regarding the local control of lung metastases, MWA is associated with a number of significant advantages.

Enhancing geospatial prediction models with feature engineering from road networks: a graph-driven approach

Traditional geospatial predictive models for property valuation have naturally relied on coordinates as well as ‘hedonic’ (internal and external) features. In particular, location-centric methods such as Geographically Weighted Regression (GWR) and Kriging have focused on intrinsic target characteristics together with distances between individual targets. However, especially in the context of heavily urbanized areas, these approaches might overlook crucial aspects arising from the underlying topological structure that presents itself in such areas. Concretely, in this work, we focus on the structure arising from the road network connecting properties. We introduce a novel though straightforward technique for feature engineering based on graphs constructed on a road network. We then extract relevant features from these and utilize those as inputs for predictive models and assess their performance benefits when used together with a variety of both well-known geospatial models and state-of-art machine learning models. To this end, we present an exhaustive experiment using four different real-life data sets across various regions and exhibiting sizes outperforming many comparative works in the field. Our findings reveal that our feature engineering approach offers significant improvements in predictive performance. Finally, we apply Shapley values as an interpretability technique to confirm the reliability and effectiveness of our approach.

High-velocity projectile penetration test and theoretical calculation of pseudo fluid penetration of calcareous sand

To explore the penetration resistance of calcareous sand media, penetration tests have been conducted in the velocity range of 200–1000 m/s using conical-nosed projectiles with a diameter of 14.5 mm. Further, a pseudo fluid penetration model applicable to the penetration of rigid projectiles in sand media is established according to the approximate flow of compacted sand in the adjacent zone of penetration. The correlation between the impedance function of projectile-target interaction and the internal friction features of pseudo fluid is clarified, and the effects of sand density, cone angle of nose-shaped projectile, and dynamic hardness on the penetration depth are investigated. The results verify the feasibility, wide applicability, and much lower error (with respect to the experimental data) of the proposed model as compared to the Slepyan hydrodynamic model.

End-to-end Temporal Transformer with Autocorrelated Attention Mechanism Augmented for Remaining Useful Life Prediction of Turbopump Bearings

Accurately predicting the remaining useful life (RUL) has become crucial for ensuring stable and safe operations for rocket engines due to the extreme working environment. However, current RUL prediction approaches based on convolution and recurrent frameworks lack effective feature extraction methods to model long-term dependencies, resulting in limited accuracy and generalizability. To address this issue, we propose an end-to-end temporal Transformer with autocorrelated attention mechanism augmented for RUL prediction of turbopump bearings. The Transformer module is adopted as the backbone of proposed framework to model long-term dependencies from the raw signals. To further enhance predictive capability, we develop a self-attention mechanism based on autocorrelation calculation. This mechanism extracts and aggregates feature representations through similarity comparison at the sub-series level. Furthermore, we utilize convolutional layers with residual links to capture internal detail features, compensating for the limitations of capturing local information. The proposed framework is evaluated through a life-cycle rocket engine bearing dataset and the experimental results demonstrate the effectiveness and superiority on RUL prediction.

Interactive three-dimensional atlas of the mineralized skeleton of the sand tiger shark Carcharias taurus.

Non-invasive computed tomography (CT) of an adult sand tiger shark Carcharias taurus Rafinesque, 1810 is used to provide an interactive three-dimensional 'general' shark (Selachimorpha) anatomy atlas. Given its post-cranial body morphology, the sand tiger shark appeared to be a well-chosen candidate and through comparison of the sand tiger shark with several other representatives of all eight established orders of sharks, we confirm that the relatively large degree of mineralization of the endoskeleton, along with the overall size, makes the sand tiger shark an ideal candidate for skeletal segmentation and construction of a skeletal atlas using conventional CT. This atlas both increases accessibility to the internal morphological features of the sand tiger shark and provides a more generalized overview of the skeletal anatomy of sharks and can aid as a supplement to destructive fresh dissection of specimens in the future and the construction of future skeletal atlases of other less mineralized sharks.

Green Infrastructure Fluctuations in Urban Agglomeration of Shanxi Province, China: Implications for Controlling Ecological Crises

The rapid urbanization process means that even moderate-sized cities can quickly become part of larger urban agglomerations, creating new urban zones. Urban Green Infrastructure (UGI) plays a crucial role in these clusters, acting as precious green spaces essential for maintaining ecological safety. This study combines fluctuation analysis based on Morphological Spatial Pattern with traditional landscape pattern analysis, comprehensively addressing the evolution of UGI in terms of quantity, characteristics, and morphology. We selected the Taiyuan-Jinzhong agglomeration as our study area, which is currently in an agglomeration process. The results demonstrated the critical role of surrounding mountains as natural ecological barrier zones. During urban agglomeration, management strategies focused on large-scale afforestation to ensure the quantity of UGI. However, this approach also led to a more clustered landscape with reduced connectivity. Additionally, linear or small-scale UGI types such as branch and islet have seen reductions over the past decade. Changes in internal morphological and complex fluctuations within UGI can harm the formation of ecological networks and potentially negatively affect biodiversity and ecological safety. The research highlights how ecological protection and urban planning policies can influence UGI fluctuations. Therefore, urban managers should not just concentrate on maintaining the quantity of UGI, but also give consideration to changes in its internal features and morphology. Before cities further agglomerate into larger urban clusters, it is crucial to address deficiencies in UGI, continuously improving type configurations and functional structures at the landscape scale. Through strategic planning of UGI, cities can mitigate ecological risks and foster sustainable urban development.

The Extraction of Roof Feature Lines of Traditional Chinese Village Buildings Based on UAV Dense Matching Point Clouds

Traditional Chinese buildings serve as a carrier for the inheritance of traditional culture and national characteristics. In the context of rural revitalization, achieving the 3D reconstruction of traditional village buildings is a crucial technical approach to promoting rural planning, improving living environments, and establishing digital villages. However, traditional algorithms primarily target urban buildings, exhibiting limited adaptability and less ideal feature extraction performance for traditional residential buildings. As a result, guaranteeing the accuracy and reliability of 3D models for different types of traditional buildings remains challenging. In this paper, taking Jingping Village in Western Hunan as an example, we propose a method that combines multiple algorithms based on the slope segmentation of the roof to extract feature lines. Firstly, the VDVI and CSF algorithms are used to extract the building and roof point clouds based on the MVS point cloud. Secondly, according to roof features, village buildings are classified, and a 3D roof point cloud is projected into 2D regular grid data. Finally, the roof slope is segmented via slope direction, and internal and external feature lines are obtained after refinement through Canny edge detection and Hough straight line detection. The results indicate that the CSF algorithm can effectively extract the roofs of I-shaped, L-shaped, and U-shaped traditional buildings. The accuracy of roof surface segmentation based on slope exceeds 99.6%, which is significantly better than the RANSAC algorithm and the region segmentation algorithm. This method is capable of efficiently extracting the characteristic lines of roofs in low-rise buildings within traditional villages. It provides a reference method for achieving the high-precision modeling of traditional village architecture at a low cost and with high efficiency.

Eye movement differences when recognising and learning moving and static faces.

Seeing a face in motion can help subsequent face recognition. Several explanations have been proposed for this "motion advantage," but other factors that might play a role have received less attention. For example, facial movement might enhance recognition by attracting attention to the internal facial features, thereby facilitating identification. However, there is no direct evidence that motion increases attention to regions of the face that facilitate identification (i.e., internal features) compared with static faces. We tested this hypothesis by recording participants' eye movements while they completed the famous face recognition (Experiment 1, N = 32), and face-learning (Experiment 2, N = 60, Experiment 3, N = 68) tasks, with presentation style manipulated (moving or static). Across all three experiments, a motion advantage was found, and participants directed a higher proportion of fixations to the internal features (i.e., eyes, nose, and mouth) of moving faces versus static. Conversely, the proportion of fixations to the internal non-feature area (i.e., cheeks, forehead, chin) and external area (Experiment 3) was significantly reduced for moving compared with static faces (all ps < .05). Results suggest that during both familiar and unfamiliar face recognition, facial motion is associated with increased attention to internal facial features, but only during familiar face recognition is the magnitude of the motion advantage significantly related functionally to the proportion of fixations directed to the internal features.

Gross pathology of brain mass lesions by intraoperative ultrasonography: a comparative study

BackgroundThe purpose of this study was to evaluate the ability of intraoperative ultrasound (IOUS) to differentiate the gross pathological features of brain mass lesions in comparison with preoperative imaging and confirmable histopathological results.ResultsA total of 365 patients were operated on for brain mass lesions removal from May 2017 to May 2022 under the guidance of intraoperative ultrasound with transducers 2.5–8 megahertz (MHZ). Ultrasound gross pathological findings were compared to the preoperative imaging and the confirmable histopathological results. Intraoperative ultrasound defined either internal or external gross pathological features of all brain mass lesions. The IOUS showed spontaneous enhancements of the brain abscess walls, which were equivalent to contrasted CT and MRI. Significantly large diameters were noted in the IOUS measurement of abscesses in comparison with CT and MRI (P = 0.001). The walls of the brain abscesses were significantly well defined in IOUS imaging in comparison with CT (P = 0.001) and equivalent to MRI. IOUS showed equivalent significance to CT and MRI in characterizing intra-parenchymal hematomas. Significantly large diameters were noted in the IOUS measurement of hematomas in comparison with CT and MRI (P = 0.001). IOUS showed significant definition of brain tumors in comparison with CT and MRI regarding tumor edge definition, tumor contours, necrosis, and cystic components (cystic definition, cystic multiplicity, cystic trabeculations, and cystic wall thickness) (P = 0.001). IOUS was equivalent to CT and MRI regarding intra-tumoral hematomas and brain edema. IOUS was equivalent to CT regarding calcification detection. The significant criteria for high-grade brain tumors versus low grade by IOUS were: P = 0.001 (necrosis, brain edema, rare calcifications, presence of cystic components, thick cystic walls, large diameter, hypo-echogenicity, and heterogeneity); P = 0.002 (cystic trabeculations); P = 0.005 (multiple cysts); and P = 0.03 (irregular contour). IOUS can characterize brain tumors and suspect specific and significant criteria for many types with great overlap.ConclusionsIntraoperative ultrasound has the ability to differentiate the gross pathological features of brain mass lesions in comparison with preoperative imaging and confirmable histopathological results.

Experimental construction of eddy real-time structure based on gradient-dependent OI in the Kuroshio-Oyashio confluence region

The internal structures of mesoscale eddies are variable due to different generation mechanism, spatiotemporal scale, and movement characteristics. Based on the principle of data assimilation, this study provided a new approach for constructing the eddy real-time internal structure using the gradient-dependent optimal interpolation (OI) method. Three eddy cases with different types and locations were selected to validate the construction results. The reliability of construction was evaluated by satellite observation, in-situ comparisons, and contrast with the numerical output. The results showed that the real-time geostrophic current characters were basically consistent with the surface and deep current (observed by satellite and acoustic doppler current profiler, ADCP). The density features exhibited by constructed results (eddy centers, mean radii, and variation tendency) were similar to the expendable conductivity-temperature-depth (XCTD) survey data and numerical output. Compared to the composite structure, the real-time structures represented more small or mesoscale fluctuations. During eddy cases evolution, the mean radii at the same depth first increased and then decreased. The horizontal scale, influence depth, and internal stability of an anticyclonic eddy with low-latitude were more powerful than the other two eddy cases owing to the domination of warm water in the Kuroshio extension region. The results indicated that the gradient-dependent OI was a hopeful technique for representing the real-time internal features during eddy evolution.

Knowledge Distillation Based on Fitting Ground-Truth Distribution of Images

Knowledge distillation based on the features from the penultimate layer allows the student (lightweight model) to efficiently mimic the internal feature outputs of the teacher (high-capacity model). However, the training data may not conform to the ground-truth distribution of images in terms of classes and features. We propose two knowledge distillation algorithms to solve the above problem from the directions of fitting the ground-truth distribution of classes and fitting the ground-truth distribution of features, respectively. The former uses teacher labels to supervise student classification output instead of dataset labels, while the latter designs feature temperature parameters to correct teachers’ abnormal feature distribution output. We conducted knowledge distillation experiments on the ImageNet-2012 and Cifar-100 datasets using seven sets of homogeneous models and six sets of heterogeneous models. The experimental results show that our proposed algorithms improve the performance of penultimate layer feature knowledge distillation and outperform other existing knowledge distillation methods in terms of classification performance and generalization ability.

Optimizing Tensile Properties and Hardness of Inconel 718 by Cold Rolling

The as-received commercial Inconel 718 material was solid solution heat treated (ST), cold-rolled (CR), and precipitation-hardened (PH) to investigate the effects of deformation on the tensile properties, hardness, and texture. Three sets of specimens (0%, 20%, and 50% CR) were ST at 1100 °C/1 h, CR, and aged (720 °C/8 h + 650 °C/8 h) for the analysis. The ultimate tensile strength (UTS), 0.2% yield strength (YS), and elongation of 50% deformed condition were 1645 MPa, 1512 MPa, and 3.8%, respectively. The 20% deformation resulted in a balanced UTS (1348 MPa), YS (1202 MPa), and elongation (11%). The contribution of precipitation hardening to the strength decreased, while the contribution of CR increased with an increasing percentage of deformation. As the level of deformation increased, the size and quantity of γ″ decreased proportionally. The CR specimens produced a high density of nano/micro twins with twin planes oriented perpendicular to the RD-ND surface. The gradient of crystal orientation and internal features of large austenitic grains were generated by their preferred rotation. The cross-slip of screw dislocations induced a complete β-fiber consisting of {110}&lt;112&gt;, {112}&lt;111&gt;, and {123}&lt;634&gt;, with doubled intensity at the higher deformation. Additionally, the specimens were highly susceptible to the twinning-induced orientation emerged by a predominant &lt;110&gt;//ND. In the as-deformed condition, an incomplete but intense α-fiber, clustered between {110}&lt;001&gt; and {110}&lt;112&gt;, was characterized. Apart from achieving the highest strength, the current work demonstrates the effects of CR on the material strength without the complex influences of δ precipitates.

Entrepreneurship Education in Interactions in English Terminale and Mastering English (High School), Cameroon EFL/ESL Textbooks: A Critical Discourse Analysis

Entrepreneurship education in the 21st century has increasingly become an important topic in the field of language teaching. EFL/ESL textbooks, in the process of teaching English, have incorporated, among themes of interest, entrepreneurship, which is helpful to transform students’ environments, improve their socio-economic status, and keep them away from poverty. However, the teaching methodology and the quality of materials selected bear questionable social practices likely to denature textbook designers’ laudable intentions. This paper is a critical discourse analysis of the teaching of entrepreneurship in two selected Cameroon English language textbooks used in the French and English subsystems of education, namely Interactions in English Terminale and Mastering English (High School). It is a sociolinguistic qualitative study which hinges on Fairclough’s (2001) three-dimensional theoretical paradigm, which holds that discourses are not language-bound but encompass internal features that (re)produce social structures, social practices, and power relations, reinforcing social inequalities and hegemonic ideologies. The findings yielded significant findings, showing that both Interactions in English Terminale and Mastering English textbooks are replications of the political, economic, and social institutions of Cameroon society and inadvertently promote gender discrimination in terms of entrepreneurship representation. The textbooks examined are too idealistic in their entrepreneurial discourses and outcomes. They fail to expose students to the realities of entrepreneurial ventures, thereby depriving them of the risks and challenges faced by the prominent entrepreneurs celebrated in their course books. This study is relevant as it contributes to the improvement of entrepreneurship pedagogy in EFL/ESL textbooks used in Cameroon classrooms while adopting realistic overtones.

Development of Multimodal Fusion Technology for Tomato Maturity Assessment.

The maturity of fruits and vegetables such as tomatoes significantly impacts indicators of their quality, such as taste, nutritional value, and shelf life, making maturity determination vital in agricultural production and the food processing industry. Tomatoes mature from the inside out, leading to an uneven ripening process inside and outside, and these situations make it very challenging to judge their maturity with the help of a single modality. In this paper, we propose a deep learning-assisted multimodal data fusion technique combining color imaging, spectroscopy, and haptic sensing for the maturity assessment of tomatoes. The method uses feature fusion to integrate feature information from images, near-infrared spectra, and haptic modalities into a unified feature set and then classifies the maturity of tomatoes through deep learning. Each modality independently extracts features, capturing the tomatoes' exterior color from color images, internal and surface spectral features linked to chemical compositions in the visible and near-infrared spectra (350 nm to 1100 nm), and physical firmness using haptic sensing. By combining preprocessed and extracted features from multiple modalities, data fusion creates a comprehensive representation of information from all three modalities using an eigenvector in an eigenspace suitable for tomato maturity assessment. Then, a fully connected neural network is constructed to process these fused data. This neural network model achieves 99.4% accuracy in tomato maturity classification, surpassing single-modal methods (color imaging: 94.2%; spectroscopy: 87.8%; haptics: 87.2%). For internal and external maturity unevenness, the classification accuracy reaches 94.4%, demonstrating effective results. A comparative analysis of performance between multimodal fusion and single-modal methods validates the stability and applicability of the multimodal fusion technique. These findings demonstrate the key benefits of multimodal fusion in terms of improving the accuracy of tomato ripening classification and provide a strong theoretical and practical basis for applying multimodal fusion technology to classify the quality and maturity of other fruits and vegetables. Utilizing deep learning (a fully connected neural network) for processing multimodal data provides a new and efficient non-destructive approach for the massive classification of agricultural and food products.

An innovative MGM–BPNN–ARIMA model for China’s energy consumption structure forecasting from the perspective of compositional data

Effective forecasting of energy consumption structure is vital for China to reach its “dual carbon” objective. However, little attention has been paid to existing studies on the holistic nature and internal properties of energy consumption structure. Therefore, this paper incorporates the theory of compositional data into the study of energy consumption structure, which not only takes into account the specificity of the internal features of the structure, but also digs deeper into the relative information. Meanwhile, based on the minimization theory of squares of the Aitchison distance in the compositional data, a combined model based on the three single models, namely the metabolism grey model (MGM), back-propagation neural network (BPNN) model, and autoregressive integrated moving average (ARIMA) model, is structured in this paper. The forecast results of the energy consumption structure in 2023–2040 indicate that the future energy consumption structure of China will evolve towards a more diversified pattern, but the proportion of natural gas and non-fossil energy has yet to meet the policy goals set by the government. This paper not only suggests that compositional data from joint prediction models have a high applicability value in the energy sector, but also has some theoretical significance for adapting and improving the energy consumption structure in China.

Combining 3D printing and injection moulding for the fabrication of polymer micro-components with internal hollow features

There is a need to cost-effectively produce polymer components with meso/micro-scale internal geometries with high replication accuracy without the use of post-processing steps. A possible process chain to produce such polymer components with internal hollow features is by combining the 3D printing (3DP) and micro-injection moulding (MIM) processes. To date, no studies were carried out to explore the feasibility of such a process chain. Consequently, this experimental study investigated the use of the 3DP lost-cores that are over-moulded using the MIM process. The first step involved the production of lost-core from a soluble polymer material where three different materials were studied: two filament-based materials (Xioneer VXL130 and AquaSys180) and one resin-based material (IM-HDT-WS). The filament-based materials were printed on an Ultimaker S5 (filament fused fabrication) and the resin-based material was printed using an Asiga Max X27 (digital light processing). In the second step, the lost core was then over-moulded with polymethyl methacrylate (PMMA) using the MIM process. After demoulding, the internal core was then dissolved using the respective dissolution method of each material to achieve a part with meso/micro scale internal features. Investigations carried out at the different stages of the process chain revealed that the best dimensional accuracy was achieved when using the IM-HDT-WS material in the 3DP of the lost-cores and their subsequent over-moulding to form the case study part internal geometry. In particular, the dimensional analysis of the replicated IM-HDT-WS lost-core geometries onto the over-moulded PMMA revealed a difference of 0% in diameter and − 3.17% in bifurcation angle of the Y1.6 channel and a difference of + 4.88% in diameter and + 11.48% in bifurcation angle of the Y0.8 channels when compared to the respective 3DP core dimensional values prior to encapsulation. However, dissolution tests revealed that the filament-based material, the Xioneer VXL130, achieved a dissolution rate of 3.5 and 4.5 h for the Y1.6 and Y0.8 channel, respectively, which was marginally faster than that of the IM-HDT-WS.

Superior supercapacitor performance with tuneable 2D/3D morphological microporous carbons of zeolitic imidazolate frameworks synthesized by recycling mother liquors

Carbonized metal–organic frameworks (MOFs)-based nanoporous carbon materials (NPCs) offer attractive activities in electrochemical energy conversion and storage (EECS) applications; however, there is the need for scalable MOFs production under reduced energy/environmental impact. This study reports a green synthesis route for model zeolitic imidazolate framework (ZIF-8) materials via recycling methanol-based mother liquors (ZIF-RMLx) under room temperature stirring and their carbonized materials (CZC-RMLx) for high-performance supercapacitors. Series of ZIF-RMLx samples produced in four recycles offer rod-/sheet-/polyhedral-like 2D/3D microstructures with tuneable internal and external framework and morphological features. Accordingly, CZC-RMLx, obtained by direct pyrolysis, with high microporosity and surface area of 1500 m2 g−1 deliver excellent capacitance values of 200–340 F g−1, compared to typical MOFs-derived or chemically activated/templated NPCs, produced via extended chemical processing. Structure-relevant and comparative performance analysis reveal insights for improved charge storage and carbonization-dependent graphitization, nitrogen-doping and microporosity-controlled capacitance characteristics in the CZC-RMLx over typical NPCs in literature. The device-level performance with a long-term durability over 21,000 cycles is demonstrated. The practical potential of CZC-RMLx is further evaluated by fabricating solid-state cells and their parallel and series circuit combinations result in overall capacitance and voltage boost to 450 F g−1 and 2.4 V, respectively.

Advancing Drug-Target Interaction prediction with BERT and subsequence embedding

Exploring the relationship between proteins and drugs plays a significant role in discovering new synthetic drugs. The Drug-Target Interaction (DTI) prediction is a fundamental task in the relationship between proteins and drugs. Unlike encoding proteins by amino acids, we use amino acid subsequence to encode proteins, which simulates the biological process of DTI better. For this research purpose, we proposed a novel deep learning framework based on Bidirectional Encoder Representation from Transformers (BERT), which integrates high-frequency subsequence embedding and transfer learning methods to complete the DTI prediction task. As the first key module, subsequence embedding allows to explore the functional interaction units from drug and protein sequences and then contribute to finding DTI modules. As the second key module, transfer learning promotes the model learn the common DTI features from protein and drug sequences in a large dataset. Overall, the BERT-based model can learn two kinds features through the multi-head self-attention mechanism: internal features of sequence and interaction features of both proteins and drugs, respectively. Compared with other methods, BERT-based methods enable more DTI-related features to be discovered by means of attention scores which associated with tokenized protein/drug subsequences.We conducted extensive experiments for the DTI prediction task on three different benchmark datasets. The experimental results show that the model achieves an average prediction metrics higher than most baseline methods. In order to verify the importance of transfer learning, we conducted an ablation study on datasets, and the results show the superiority of transfer learning. In addition, we test the scalability of the model on the dataset in unseen drugs and proteins, and the results of the experiments show that it is acceptable in scalability.