Search

It is a rapid and irresistible developing trend that the traditional classroom learning mode is transformed into an online learning mode. In the online learning process, it is important to achieve effective voice separation and provide a good lecture experience for students. In this paper, a robust and automatic method for the recognition of speech categories (teacher speech, student speech, silence speech and overlapping speech) is presented. The system, which is part of artificial intelligence in education (AIED), is a hybrid speech processing system that includes GMM modeling, GMM clustering, silence processing and overlapping processing. The online learning environment we are based on is not a customized environment that defines the roles of teachers or students, but a more casual, broader and more general online environment. The method proposed and designed based on this situation has better adaptability and scalability. The experimental results show that our automatic detection of speech categories fares well when compared to the results from professional coders. [ABSTRACT FROM AUTHOR]

Carbon fiber reinforced polyamide-6 (CF/PA-6) composites have been widely applied in automobile, aerospace, and biomedical industries for their high mechanical properties, high thermal resistance and recyclability. On the purpose of finding ways to improve the interfacial properties, the investigation of the nanostructure and nanomechanical properties of the interphase in CF/PA-6 composites were essential. In this study, MD simulation was carried out to show the interfacial formation and nanostructure of the CF/PA-6 composite model directly at the atomic level and compute the radial distribution function, interfacial energy, total energy. Then the nanomechanical properties of the CF/PA-6 composite, such as interfacial thickness, interfacial modules, interfacial adhesion, were investigated by AFM PF-QNM model. The changes of the radical distribution function and energies over the MD simulation time indicated that the PA-6 chains adsorbed and then regularly folded on the CF surface, displaying the interfacial crystallization of the CF/PA-6 composite model. What stood out in the AFM PF-QNM tests were the abrupt decreasing of the interfacial modulus and the sharp increasing of the interfacial adhesion from those of the carbon fiber to those of the PA-6. The average interfacial thickness of the CF/PA-6 composite was 72 nm. Consistent with the simulation results, the interfacial properties were distinct from the properties of the carbon fiber and PA-6, owning to the adsorption and orderly folding of the PA-6 chains on the CF surface and the changes of the RDF and energies. [ABSTRACT FROM AUTHOR]

Emotion Recognition in Conversation (ERC) is a task aimed at predicting the emotions conveyed by an utterance in a dialogue. It is common in ERC research to integrate intra-utterance, local contextual, and global contextual information to obtain the utterance vectors. However, there exist complex semantic dependencies among these factors, and failing to model these dependencies accurately can adversely affect the effectiveness of emotion recognition. Moreover, to enhance the semantic dependencies within the context, researchers commonly introduce external commonsense knowledge after modeling it. However, injecting commonsense knowledge into the model simply without considering its potential impact can introduce unexpected noise. To address these issues, we propose a dialogue emotion model based on local–global context encoder and commonsense knowledge fusion attention. The local–global context encoder effectively integrates the information of intra-utterance, local context, and global context to capture the semantic dependencies among them. To provide more accurate external commonsense information, we present a fusion module to filter the commonsense information through multi-head attention. Our proposed method has achieved competitive results on four datasets and exhibits advantages compared with mainstream models using commonsense knowledge. [ABSTRACT FROM AUTHOR]

The Causal Emotion Entailment (CEE) task aims to extract all potential pairs of emotions and corresponding causes from the unannotated emotion document in the conversational context. Most existing methods to solve CEE task follow a two-stage pipeline framework, in which the first stage is to identify emotional clauses and cause clauses and extract clause representation,separately. And in the second stage is to construct the final emotion and cause pairs. However, they ignore the effect of the distance between clauses on emotion-cause pair matching. Here, we construct a joint framework with Window Transformer to handle this problem. The pre-trained BERT and RoBERTa are used as the text encoder to generate a local representation of clauses in a given document. Meanwhile, we feed it into 2D Window Transformer to make the clause representation sensitive to the context within the Window and to obtain the dependencies between clauses. At the same time, the document ranks the candidate clauses to extract causal emotion entailments, which enhances the representation of clause pairs (emotion pairs and cause pairs) by kernel-based relative position embedding. Experimental results indicate that the framework acquires state-of-the-art results on the benchmark dataset. [ABSTRACT FROM AUTHOR]

Multimodal emotion analysis has become a hot trend because of its wide applications, such as the question-answering system. However, in a real-world scenario, people usually have mixed or partial emotions about evaluating objects. In this paper, we introduce a fuzzy temporal convolutional network based on contextual self-attention (CSAT-FTCN) to address these challenges, which has a membership function modeling various fuzzy emotions for understanding emotions in a more profound sense. Moreover, the CSAT-FTCN can obtain the dependency relationships of target utterances on internal own key information and external contextual information to understand emotions in a more profound sense. Additionally, as for multi-modality data, we introduce an attention fusion (ATF) mechanism to capture the dependency relationship between different modality information. The experimental results show that our CSAT-FTCN outperforms state-of-the-art models on tested datasets. The CSAT-FTCN network provides a novel method for multimodal emotion analysis. [ABSTRACT FROM AUTHOR]

Most existing approaches for cross-subject electroencephalogram (EEG) emotion recognition learn the universal features between different subjects with the neurological findings. The performance of these methods may be sub-optimal due to the inadequate investigation of the relationships between the brain and the emotion. Hence, in case of insufficient neurological findings, it is essential to develop a domain adaptation method for EEG data. In this paper, we propose a generator-based domain adaptation method with knowledge free (GDAKF) mechanism for the cross-subject EEG emotion recognition. Specifically, the feature distribution of the source domain is transformed into a feature distribution of the target domain via adversarial learning between the generator and the discriminator. Additionally, the transformation process is constrained by the EEG content regression loss and emotion information loss to maintain the emotional information during the feature alignment. To evaluate the effectiveness and performance of GDAKF, many experiments are carried out on the benchmark dataset, DEAP. The experimental result shows that GDAKF achieves excellent performance with 63.85% mean accuracy in low/high valence, which shows that the proposed method is comparable to the EEG cross-subject emotion recognition methods in the literature. This paper provides a novel idea for addressing cross-subject EEG emotion recognition, and it can also be applied to cross-session and cross-device emotion recognition tasks. [ABSTRACT FROM AUTHOR]

The design of genetic operators is absolutely one of the core work of evolutionary algorithms research. However, the essence of the evolutionary algorithms is that a lot of algorithm design is based on the manual result analysis, summarize, refine, feedback, and then, the algorithms are designed adaptively and correspondingly. This kind of design scheme needs artificial statistics and analysis of large amounts of data, which greatly increases the burden of the designers. To solve this problem, an evolutionary algorithm framework based on genetic operator automatic design is proposed in this paper. In the first step, Gene Expression Programming and Differential Evolution methods are combined together and used to design the genetic operators automatically and adaptively, this hybrid method can not only explore solutions in problem space for the problem solving as most classical evolutionary algorithms do, but also generate genetic operators automatically in operator space for the proper operators extraction and selection related to the evolutionary algorithms. In the second step, the designed operators are adopted into the typical evolutionary algorithms to verify the performance and the result shows that the new designed genetic operator is superior to or at least equivalent to some existing DE variants in a set of classical benchmark functions. More importantly, this paper is not aimed at designing high performance algorithms, but to provide a new perspective for algorithms designing, and to provide a reference scheme for the machine algorithms designing. [ABSTRACT FROM AUTHOR]

Ultrathin, flexible and highly conductive materials that possess excellent electromagnetic interference (EMI) shielding performance are greatly needed, especially for the fabrication of stretchable shielding materials in practical applications such as wearable and foldable electronics. Graphene oxide (GO) sheets are modified with ethylenediamine to prepare cross-linked graphene films using a pressure-assisted self-assembly technique. FTIR and XPS results demonstrate that amine monomers are chemically bonded to GO sheets, with simultaneous reduction of GO sheets. After thermal annealing and followed with compression, the 6.6-μm-thick nitrogen-doping graphene film (rGO-EDA-2) is obtained with ultrahigh electrical conductivity of 8796 S cm−1. The excellent electrical conductivity is mainly attributed to nitrogen-doping effect, defects repair during chemical functionalization and removal of oxygenated groups. Ultrahigh electrical conductivity, multilayer structure and modified electronic structure with nitrogen doping lead to outstanding shielding performance for the rGO-EDA-2 film, with excellent shielding effectiveness (SE) of 58.5 dB and the specific SE/thickness of 43902 dB cm2 g−1, respectively. By fixing the rGO-EDA-2 film on the pre-stretched wavy substrate, the stretchable shielding composite is obtained, with constant EMI SE of 56.3 dB after repeated stretching. The pre-stretched wavy substrate allows the multilayer graphene film to achieve wavy structure after strain release, which is capable of bearing tensile strain up to 32.6%. This study could be significant in the applications of stretchable and wearable electronic devices. [ABSTRACT FROM AUTHOR]

To evaluate the performance and reliability of a new type of epoxy resin adhesive, the dynamic mechanical analysis (DMA) testing was performed at different levels of temperature and frequency after hydrothermal aging. The results of the DMA and the thermodynamic analysis that the performance of the adhesive has little change after hydrothermal aging for 30 days meant that the resistance to hydrothermal aging of this adhesive was excellent. In addition, the time-temperature equivalence principle and time-aging time equivalence principle were proposed here to offer additional insights into the dynamic mechanical performance of the newly developed adhesive. Moreover, an effective method by which the generalized curve of definite aging time in certain conditions of definite temperature and humidity can be obtained was introduced.Graphical abstract [ABSTRACT FROM AUTHOR]

User attention state recognition when interacting with a monitor or undertaking a specific task represents a crucial issue in many domains and applications, such as e-learning, driving and network video conferences. However, for the consideration of convenience and practicality in these situations, it is inescapable and necessary to develop a noncontact and automatic monitoring system to analyze, recognize and predict what kind of attention state in individuals during a task execution without delay. The elaborated technique presented here has achieved efficient cardiac pulse estimation based on the face image captured by the noncontact, automatic and webcam-based measurement method. After collecting cardiac pulse signals, various features extraction methods are presented to obtain key features from the raw data which is related to the attention state or not. The experiment result shows that it is possible to estimate humans attention state based on the technique presented here. The proposed technique may be useful for monitoring person for the purpose of health care, psychological testing, online learning or security, etc. [ABSTRACT FROM AUTHOR]

For the improvement of economic and environmental performance of bone drilling operations, the optimum value of the spindle speed and feed rate must be known. This is because the input spindle speed is an important factor of energy consumption and with its approximate value known in optimization of output characteristics of bone drilling operations may result in significant energy savings. Optimization of multi-output characteristics of the bone drilling process is possible, if the relationships between the outputs and the inputs are known. Therefore, this study forms the strong basis for development of the models for the three output characteristics (maximum temperature, maximum force and maximum average surface roughness) for the bone drilling operation performed on the bovine bone. Experimental studies are conducted to measure these three outputs based on the spindle speed and feed rate. The validation of the formulated models is done based on the root-mean-square error, coefficient of determination, relative error, multi-objective error and mean absolute percentage error. The relationships between the three outputs and the inputs are further revealed by the 2-D analysis on the models. The findings from these relationships can be used for the predictive monitoring the bone drilling operation. The work concludes with discussion of environmental implications arising from the current study. [ABSTRACT FROM AUTHOR]

Acute hypotensive episodes (AHE) is a critical event in the ICU as it can cause multiple organ failure and thus has severe implications on the mortality risk. Timely and rapid intervention is necessary to save a patient's life. For the sake of this purpose, this paper presents a methodology to predict AHE for ICU patients based on mean arterial pressure time series. First, the empirical mode decomposition method is used to calculate complex MAP series. Then the original and intrinsic mode functions of MAP series are transformed for the feature extraction with the methods of probabilistic distribution and statistics. Finally, the multi genetic programming is presented to build the classification models for the prediction of AHE. The methodology is applied to the dataset from Multi-parameter intelligent monitoring for intensive care. The achieved accuracy of the proposed methodology is 82.92 % in the training set and 79.93 % in the testing set with the sixfold cross-validation in the total 2866 patient's records. Simulation results revealed the reasonable forecasting accuracy in prediction of AHE in 1 h forecasting window. [ABSTRACT FROM AUTHOR]

The Yunmengshan metamorphic core complex in the middle part of the Yanshan Fold and Thrust Belt records crust extension processes of the eastern North China Craton during its peak destruction. Development of the metamorphic core complex was controlled by the generally NNE-striking Dashuiyu Shear Zone. The shear zone dips SE and becomes shallower NE-wards, leading to exposures of a ductile shear zone in the southern and middle parts and brittle faults in the northern part. Exposure structures, microstructures, and quartz C- axis fabrics indicate that the ductile shear zone belongs to an extensional shear zone with a top-to-the-SE shear sense. Deformation temperatures of 300-520 °C suggest a mid-crustal origin for the ductile shear zone. A ductile deformation belt in the footwall of the shear zone is only as wide as 1-3 km, indicating no widespread mid-crustal ductile flow in the region during the deformation. Zircon U-Pb dating of dykes and plutons as well as hornblende and biotite Ar/Ar dating demonstrate that the metamorphic core complex originated at 135 Ma and experienced intense shearing of the Dashuiyu Shear Zone, development of the supradetachment basins, and synkinematic intrusion during 135-125 Ma. The metamorphic core complex was subjected to rapid exhumation during 125-114 Ma when the Dashuiyu Shear Zone suffered continuous activity and passive doming. The shear zone and its hanging wall were cut or replaced by a series of brittle faults when they were uplifted to a brittle regime, showing that exhumation took place in continuous extensional activities. The metamorphic core complex turned into slow exhumation in an extensional regime in the following latest Early Cretaceous. The evolution history suggests that the Yunmengshan metamorphic core complex was developed by the rolling-hinge model, a common formation mechanism for intraplate metamorphic core complexes in the North China Craton, under the continuous NW-SE extension during the Early Cretaceous (135-100 Ma). [ABSTRACT FROM AUTHOR]

The development of a simple sensor (9NL27-Zn) based on DNAzyme and PCR and aimed at the detection of low concentrations of zinc (II) ions is described. A specific Zn(II)-dependent DNAzyme (9NL27) with DNA-cleaving activity was employed. In the presence of zinc (II), the DNAzyme hydrolyzed DNA substrate into two pieces (5′ and 3′ fragments), forming 3′-terminal hydroxyl in the 5′ fragment and 5′-phosphate in the 3′ fragments. Subsequently, the 5′ fragment left the DNAzyme and bound a short DNA template. The 5′ fragment was used as a primer and extended a single-stranded full-length template by Taq polymerase. Finally, this full-length template was amplified by PCR. The amplified products had a quantitative relationship with Zn(II) concentration. Under our experimental conditions, the DNA sensor showed sensitivity (10 nM) and high specificity for zinc ion detection. After improvement of the DNA sensor, the detection limit can reach 1 nM. The simple DNA sensor may become a DNA model for the detection of trace amounts of other targets. [Figure not available: see fulltext.] [ABSTRACT FROM AUTHOR]

In vacuum infusion molding process (VIMP), it is difficult to manufacture a composite part with small dimensional tolerance, since the upper mold for the process is flexible. In this study, the static and cyclic compaction responses of five kinds of fabrics were experimentally studied under real VIMP conditions, with the effects of compaction pressure, compaction time, compaction cycle and number of the fabric layers. The static and cyclic compaction responses of the all fabrics follow different power law models and the resulting fiber volume fraction and relaxation factor increase with the number of layers. Although the resulting fiber volume fraction increases with the layer numbers, change of the fiber volume fraction of the composite parts with 10 layers to 100 layers of the all fabrics is less than 2.5%. The thickness of the composite part was monitored and measured using micrometer gauges, and the effects of processing parameters on the final thickness of part was investigated. The part thickness varies as a function of spatial coordinates and time during pre-filling, filling and post-filling stages in VIMP. The variation and the final value of the part thickness would be significantly affected by the processing parameters. Statistical results show that the final part thickness is equivalent to the thickness of the dry preform under the 0.08 MPa vacuum compaction pressure in VIMP. The difference between the fiber volume fraction of the final part and that of the dry preform is 2% ~ 5.7%. [ABSTRACT FROM AUTHOR]

The chemorheological behaviors of a low viscosity epoxy resin system (Huntsman 1564/3486) for vacuum infusion moulding process (VIMP) were studied with viscosity experiments. The dual-Arrhenius rheological model and the engineering viscosity model were established and compared with the experimental data. The result showed that the viscosity in the earlier stage calculated by dual-Arrhenius model were smaller than the experimental data, while the data calculated by the engineering model were larger. Combining the two models together can predict the rheological behaviors of the resin system in a more credible manner. The processing windows of the resin system for VIMP were determined based on the two models. The optimum processing temperature is 30-45 °C. [ABSTRACT FROM AUTHOR]

A printer assembly consisting of a pen, wiper with two blades, and wiper blade holder is investigated in this study. The two wiper blades press onto the pen surface via proper dimension interference, while wiping across the pen at a certain speed. A finite element model (FEM) was set up with ANSYS-LSDYNA. The hyper-elastic rubber wiper blades are described by Christensen’s model with n = 1 for the strain energy density function. Dynamic contact forces between the wiper blades and the pen, number of contacts, and the duration of the wiper blades jumping over the pen are numerically simulated. Effects of the mechanical properties of the rubber materials, i.e., the coefficient of the Christensen’s model, C10, on the dynamic contact performance are investigated. The wiper blades are found to periodically touch and jump off from the pen. For the contact between the front blade and the pen, results show that the average contact forces and the number of the contact will increase with C10, and for the back wiper blade there exist the same conclusions, except that the jumping duration of the back wiper blade does not decrease steadily, but shows a minimum. These results suggest that a new design with two curved sides of the blade tip could be beneficence. [ABSTRACT FROM AUTHOR]