Testing Stage Research Articles

Deep Learning and High-Resolution Anoscopy: Development of an Interoperable Algorithm for the Detection and Differentiation of Anal Squamous Cell Carcinoma Precursors—A Multicentric Study

High-resolution anoscopy (HRA) plays a central role in the detection and treatment of precursors of anal squamous cell carcinoma (ASCC). Artificial intelligence (AI) algorithms have shown high levels of efficiency in detecting and differentiating HSIL from low-grade squamous intraepithelial lesions (LSIL) in HRA images. Our aim was to develop a deep learning system for the automatic detection and differentiation of HSIL versus LSIL using HRA images from both conventional and digital proctoscopes. A convolutional neural network (CNN) was developed based on 151 HRA exams performed at two volume centers using conventional and digital HRA systems. A total of 57,822 images were included, 28,874 images containing HSIL and 28,948 LSIL. Partial subanalyses were performed to evaluate the performance of the CNN in the subset of images acetic acid and lugol iodine staining and after treatment of the anal canal. The overall accuracy of the CNN in distinguishing HSIL from LSIL during the testing stage was 94.6%. The algorithm had an overall sensitivity and specificity of 93.6% and 95.7%, respectively (AUC 0.97). For staining with acetic acid, HSIL was differentiated from LSIL with an overall accuracy of 96.4%, while for lugol and after therapeutic manipulation, these values were 96.6% and 99.3%, respectively. The introduction of AI algorithms to HRA may enhance the early diagnosis of ASCC precursors, and this system was shown to perform adequately across conventional and digital HRA interfaces.

Effects of Maitland mobilization technique on upper extremity function in stroke survivors with spasticity: An experimental study.

The recovery of upper limb function is of great significance for stroke patients to regain their self-care ability, yet it is still a difficult point in clinical practice of neurological rehabilitation. This study aimed to investigate the effect of Maitland joint mobilization technique on the recovery of upper extremity function in patients with spasticity after stroke. From August to December 2023, 71 patients with upper extremity flexor spasm after stroke were recruited and randomly divided into experimental group (n = 35) and control group (n = 36). The control group was given conventional rehabilitation treatment, while the experimental group was treated with Maitland mobilization technique treatment of upper extremity joints on the basis of the control group. The experiment lasted for 8 weeks. Participants of the 2 groups were observed for Fugl-Meyer motor assessment-upper extremity (FMA-UE), box and block test (BBT) and Brunnstrom stage, modified Ashworth scale (MAS), and functional independence measure (FIM) at pre- and post-8 weeks study. There was no significant difference in gender distribution, hemiplegic side, diagnosis, past history, age, duration, body mass index, and mini-mental state examination between the 2 groups (P > .05). After 8 weeks of intervention, both groups showed significant improvement in FMA-UE, Brunnstrom stage, BBT, FIM, and MAS of the shoulder (P < .05); however, there was no significant change in MAS of the elbow, wrist, and finger joints (P > .05). The posttreatment values showed a significant improvement in FMA-UE, BBT, and FIM in the experimental group compared to the control group. Comparing the changes in pretreatment and posttreatment, FMA-UE, BBT, and FIM in the experimental group were significantly improved compared with those in the control group (P < .05). Maitland joint mobilization can improve the motor function of upper extremity and the spasticity of shoulder joint complex in patients with stroke.

Comparative study for the performance of pure artificial intelligence software sensor and self-organizing map assisted software sensor in predicting 5-day biochemical oxygen demand for Kauma Sewage Treatment Plant effluent in Malawi

Introduction: Modeling plays a crucial role in understanding wastewater treatment processes, yet conventional deterministic models face challenges due to complexity and uncertainty. Artificial intelligence offers an alternative, requiring no prior system knowledge. This study tested the reliability of the Adaptive Fuzzy Inference System (ANFIS), an artificial intelligence algorithm that integrates both neural networks and fuzzy logic principles, to predict effluent Biochemical Oxygen Demand. An important indicator of organic pollution in wastewater.Materials and Methods: The ANFIS models were developed and validated with historical wastewater quality data for the Kauma Sewage Treatment Plant located in Lilongwe City, Malawi. A Self Organizing Map (SOM) was applied to extract features of the raw data to enhance the performance of ANFIS. Cost-effective, quicker, and easier-to-measure variables were selected as possible predictors while using their respective correlations with effluent. Influents’ temperature, pH, dissolved oxygen, and effluent chemical oxygen demand were among the model predictors.Results and Discussions: The comparative results demonstrated that for the same model structure, the ANFIS model achieved correlation coefficients (R) of 0.92, 0.90, and 0.81 during training, testing, and validation respectively, whereas the SOM-assisted ANFIS Model achieved R Values of 0.99, 0.87 and 0.94. Overall, despite the slight decrease in R-value during the testing stage, the SOM- assisted ANFIS model outperformed the traditional ANFIS model in terms of predictive capability. A graphic user interface was developed to improve user interaction and friendliness of the developed model. Integration of the developed model with supervisory control and data acquisition system is recommended. The study also recommends widening the application of the developed model, by retraining it with data from other wastewater treatment facilities and rivers in Malawi.

Intracardiac echocardiography guided anatomical ablation of the arcuate ridge for drug refractory inappropriate sinus tachycardia.

Inappropriate sinus tachycardia (IST) is a common condition with frequently not tolerated beta-blockers or ivabradine and a high rate of complication in ablation strategy; we describe an alternative anatomical approach of sinus node (SN) modulation. This retrospective study describes a case series of 6 patients from two centers diagnosed with symptomatic IST undergoing SN ablation. The mean age was 40.6 ± 13.9 years; five of the six patients were female, 100% of patients reported heart palpitations, and 66% reported dizziness, the average heart rate(HR) on a 24-h Holter was 93.2 ± 7.9 bpm. HR during the first stage of a stress test using a standard Bruce protocol was 150 ± 70 bpm, The average HR on 24-h Holter postablation was 75 ± 5.6 bpm, the sinus rate HR during stage 1 of a Bruce protocol exercise stress test was 120 ± 10 bpm. This is the first case series reporting the acute and long-term results of a novel anatomical approach for SN modulation to treat IST targeting the arcuate ridge(AR) under intracardiac echography(ICE) guidance. The novel anatomic ICE-guided catheter ablation approach aimed to identify the earliest activation at the AR with an extension of RF lesions towardits septal region seems effective and safe to modulate the SN in symptomatic patients with IST refractory to medical treatment.

Rancang Bangun Game Adventure 3D Edukasi Sampah Organik dan Non-Organik

This research aims to design and develop an Educational 3D Adventure Game as an informative tool for organic and non-organic waste. The development method employed is MDLC (Multimedia Development Life Cycle), encompassing Concept, Design, Material Collecting, Assembly, Testing, and Distribution stages. The game was successfully created for the Android platform using Unity3D.The testing results, involving 22 respondents from Gedang Rowo Village, demonstrate a positive response with an approval rate of 90%. Media experts' evaluation reached 68%, while content experts provided a 100% approval rate. These findings indicate that "TrashHero" has effectively become an engaging and interactive information medium regarding organic and non-organic waste.Utilizing the Likert scale, the conclusion drawn from the testing affirms that "TrashHero" has achieved a remarkably high level of success. Therefore, this game is considered an effective educational tool for conveying information about waste while providing an interactive experience for users.

Media Pembelajaran Hewan Penghasil Listrik dengan Pemanfaatan Teknologi Augmented Reality untuk Siswa SMP

Augmented reality technology is a virtual object that can provide information to help users carry out work related to this technology. The learning process applied to students at SMPN 1 Praya Tengah with the theme of introducing electricity-producing animals with a learning method using printed books and verbal explanations using a whiteboard as a learning medium so that students cannot see pictures of electricity-producing animals. This research is aimed at producing more interactive learning media in introducing electricity-producing animals by using augmented re-ality technology at SMPN 1 Praya Tengah to increase interest in learning. The methodology used is the MDLC method which uses 6 stages, namely the concept, design, material collection, manu-facturing, testing and distribution stages which are implemented in learning media. The results obtained are interactive learning media on how to recognize electricity-producing animals using augmented reality which can be run on desktop-based computers. The conclusion obtained from this research is that the application of a learning media application to introduce electrici-ty-producing animals using augmented reality technology for students at SMPN 1 Praya Tengah can help students learn about electricity-producing animals in real time or in the form of 3-dimensional objects and animations. The test results used a test scale, namely a Likert scale with a value of 45.25 which can be categorized as Strongly Agree.

Modeling of ordinal longitudinal accelerometer data

Improvements in accelerometer technology has led to new types of data on which more powerful predictive models can be built to assess physical activity. This paper implements an ordinal random forest model with recursive forecasting to take into account the ordinal, longitudinal nature of responses. The data comes from 28 adults performing activities of daily living in two visits, while wearing accelerometers on the ankle, hip, right and left wrist. The first visit provided training data and the second testing data so that an indepen- dent sample, cross-validation approach could be used. For this type of data, prior responses are not available at the testing stage or in practice. However, recursive forecasts can be made with prior predictions in place of lagging responses on models which were built to use lagging responses as explanatory variables. Models are fit to account for multiple time series, with different time series for each participant in the study. We found that ordinal random forest, when the time series is taken into account, produces better accuracy rates and better linearly weighted kappa values than both ordinary ordinal forest and random forest. On the testing set, the lowest error rates were produced by the ankle (28.0%), followed by the left wrist (28.7%), hip (28.9%) and then the right wrist (30.2%) using the best performing decision model for a four-activity level response. In addition, linearly weighted kappa values indicated substantial agreement. The approach of this work can be adapted to other types of longitudinal ordinal models for improvements in modeling techniques. Journal of Statistical Research 2022, Vol. 57, No. 1-2, pp.1-19

Technological Analysis of the Production of Nickel-Containing Composite Materials

The article presents the results of obtaining a composite material by sintering nickel-containing raw materials mixed with carbon-containing materials, namely using coke and semi-coke. The sintering process was performed at a charge layer height of 240 mm and the temperature of the lower layer was T = 1200 °C. The results of the sieve analysis showed (a fraction of 10 mm) that the yield of a suitable composite material using coke was 68.3% and with semi-coke 67.0%. The average nickel and chromium content in the composite materials was 1.42% and 3.07%, accordingly. As a result of determining the strength characteristics of the obtained composite materials with various reducing agents by dropping from a height of 2 m onto a steel pallet, it was found that the obtained composite materials have high mechanical properties in terms of strength of 81% and 89.2%. The results of the elemental composition at the studied points and the thermal analysis of the studied composite material are presented. The mineralogical composition of the composite material is presented in the form of serpentine and nontronite, and the empty rock is made of quartz and talc. The activation energy of thermal analysis by the method of non-isothermal kinetics were calculated. The results of experiments on the production of composite materials from nickel-containing raw materials will be recommended for obtaining the optimal composition of composite materials at the stage of pilot tests and industrial development of the developed technology for processing nickel ores of the Republic of Kazakhstan. For the processing of nickel-poor nickel ores, it is of great importance to obtain optimal technological and technical and economic indicators that ensure low cost of nickel in the resulting product.

ИЗУЧЕНИЕ ВАЛИДАЦИОННЫХ ХАРАКТЕРИСТИК МЕТОДИК ИСПЫТАНИЯ НА ПОДЛИННОСТЬ И КОЛИЧЕСТВЕННОГО ОПРЕДЕЛЕНИЯ АЦИЗОЛА В СТОМАТОЛОГИЧЕСКОМ ГЕЛЕ

In the quality assurance system for medicinal products, an important role is played by the procedure for validation of analytical methods at the stages of development, testing, technology and serial production. The need for validation at the stage of drug development is due to the fact that it is possible to promptly identify the shortcomings of the method and significantly improve it. The purpose of this study is to validate methods for testing «The authenticity» and «Quantitative determination» of acyzole in dental gel. Materials and methods. To achieve this goal, an active pharmaceutical substance of pharmacopoeial quality was used. When developing methods for testing for “Authenticity,” a reaction with oxalic acid to the zinc cation determined in the composition of acyzole was used. For the quantitative determination of acyzole, the complexometric method was used. The objects of the study were three series of gel samples. Results and discussion. As a result of the research, a method for testing «Authenticity» was developed; it was found that the proposed method is characterized by a negative analytical signal on model mixtures free of the analyte component and placebo, and a positive analytical signal on model mixtures of various compositions containing the analyte component. The validation characteristics of methods for the quantitative determination of acyzole in the gel were studied. positive results were obtained. The proposed methods can be used for inclusion in the regulatory documentation for the developed gel. Conclusion. When assessing the validation characteristics of the proposed methods, it was established that the methods for testing the «Authenticity» and «Quantitative determination» of acyzole in the gel are specific, the methods of quantitative determination are characterized by accuracy and repeatability, a linear dependence in the analytical range of ±20% of the declared amount of acyzole, which allows them to be used for a reliable assessment quality of the developed dental gel.

Penerapan Metode Human Centered Design pada Perancangan Sistem Informasi Curhat Online Berbasis Web

Mental health is a very worrying health case in recent times. The number of cases that have hit the young generation of our nation is a serious threat. Pressures and challenges are often major factors in mental health crises and lack of support and places to tell stories to make a person keep everything in his mind that can hurt his mind and mentality. In this case, the research aims to overcome the problem through the implementation of a website-based online vent information system designed as a means of online vent using a Human-Centered Design approach. This Human-Centered Design approach involves a stage of user context specification, user needs, system design, and system testing using the Blackbox testing method and WebQual method. The main objective of this information system is to provide a safe and structured platform for individuals to express their feelings and emotions comfortably and safely, as well as provide support to reduce the level of mental and mental stress among the young generation of Indonesia.

Perancangan Sistem Emergency Call untuk Pengendara Bermotor Berbasis Mobile

The progress of land transportation, especially motorized vehicles, has a great influence on people's welfare. Currently motorized vehicles have a large contribution to the development of transportation in Indonesia. The increasing number of vehicles always results in problems such as flat tires, running out of petrol or breaking down and the number of these incidents always increases. When problems occur on the highway, drivers always have difficulty finding locations and information about relevant places in order to ask for help with the problems they are facing. This research aims to enable motorized vehicle users to have an Android-based mobile application that is useful for providing location information and related parties to speed up the rescue process. The application that will be developed has a emergency call feature for users to service providers or relevant locations according to the driver's problem. Not only that, this system is also a business platform for those who want to join the merchant feature in it. In this mobile application study, the author will utilize the Google Maps API so that he can search and recommend the nearest place for motorists to seek help. This research was carried out from the analysis, design, implementation, to testing stages.

Effect of Argon Flow Rate on Power Consumption of a 120‐t Ladle Furnace

Effect of Argon Flow Rate on Power Consumption of a 120‐t Ladle Furnace

Experimental study on I/II/III mixed mode fracture characteristics of a combined rock mass under creep loading

I/II/III mixed mode fractures of intersecting joint fissures often occur in natural rock masses, and jointed rock masses are prone to rockbursts in deep underground engineering when subjected to long-term crustal stresses. However, most studies of the mechanical mechanisms of these intersected joints have been conducted by simplifying two-dimensional joint model tests. Furthermore, the fracture mechanisms of two-dimensional intersected joints under tension and compression are completely different from those of three-dimensional joints. This paper presents a novel prefabricated specimen with combinations of intersecting joints capable of detecting the failure behaviours of rock I/II/III mixed mode fractures under creep loading. Uniaxial compression and multistage creep tests are performed on prefabricated sandstone specimens with intersecting joints of 0°/0°, 0°/30°, 0°/60°, and 0°/90°. The experimental results show that with the increase in the number of prefabricated intersecting joints, the uniaxial compressive strength and elastic modulus values of the sandstone specimens gradually decrease. In addition, the sandstone specimens experience relatively few AE events and minor axial strain variations in the first creep stage and the second creep stage of the multistage creep test. The axial strain increases sharply due to the sharp increase in the number of AE events in the third creep stage. The 0°/60° sandstone specimen undergoes accelerated creep failure, resulting in mixed X-shaped tensile‒shear rupture. The RA value is high based on the quantification of the creeping cracks using the acoustic emission parameters of the rise angle (RA) and average frequency (AF). The AF values of the 0°/0°, 0°/30°, and 0°/90° sandstone specimens are high. The experimental results show that a larger joint intersection angle leads to greater mutual restraints and greater effects of prefabricated crack propagation in the rock specimens, thus increasing the final failure strength. Finally, based on the acoustic emission count, a characteristic variable D suitable for characterizing the creep damage evolution of a joint rock mass is established. The findings of this paper can facilitate an effective understanding of the creep effect of I/II/III mixed mode fracture and its micromechanism. The research results will have a certain reference value for the detection and risk mitigation of instantaneous and time-delayed rockbursts.

Effectiveness of strengthening exercises on positive thinking skills, psychological wellness and mental strength in athletes with a meniscus-ripped knee

This study aims to examine the influence of positive thinking, psychological well-being, and mental toughness on the recovery process following strengthening exercises in individuals diagnosed with a meniscus tear in the knee. The research cohort comprises 32 athletes diagnosed with a meniscus tear through Magnetic Resonance Imaging (MRI). Of these participants, 26 are male, and 6 are female, all aged above 18, engaged in amateur sports across diverse disciplines, and having sustained injuries during athletic activities. The participants were randomly assigned to two groups: Group A (experimental, n=16) and Group B (control, n=16). The study adopts a mixed model incorporating experimental, survey, and relational methodologies. The experimental model utilizes the experimental method with pretest-follow-up test and posttest control group design. Over an 8-week period, participants engage in exercises with each training session comprising 10 repetitions conducted twice daily. Assessments using the Positive Thinking, Psychological Well-Being, and Mental Resilience scales occur during the Pre-Test, Follow-up Test, and Post-Test stages throughout the study duration. Analysis of the research results reveals mean scores for positive thinking and psychological well-being scales above the medium level, whereas mental toughness scale scores approximate the middle level. The study concludes a moderate, yet statistically insignificant, positive correlation among positive thinking, psychological well-being, and mental toughness scale scores. Significantly, participants, guided by a physical therapist, exhibit a faster recovery rate compared to the control group. Consequently, it is inferred that the incorporation of psychological support alongside physical exercises contributes to an accelerated recovery process.

DIGITALISASI SISTEM INFORMASI ANJUNGAN PENDAFTARAN MANDIRI (APM) RAWAT JALAN BERBASIS WEBSITE MENGGUNAKAN METODE RAPID APPLICATION DEVELOPMENT (RAD) DI RSU HAJI MEDAN

In order to achieve faster and more accurate hospital services, hospitals need to establish and have an electronic based website. The problem that still exists at Medan Haji General Hospital is related to the absence of the use of a website based outpatient self-registration platform (APM) information system. This is because officers still lack understanding of the importance of website based information systems and the limitations of an IT person to design the application. The purpose of this research is to design a website based information system for self registration booths (APM) at the Haji General Hospital in Medan. The method in this study uses the Rapid Application Development (RAD) model. The results that were concluded from the respondents were to find out whether or not there were deficiencies in the website that had been designed, and to find out whether the website system that had been designed was accepted or not by the Medan Haji Hospital. Researchers created a website design using Microsoft Office Visio 2013 to create Entity Relationship Diagrams (ERD) and Data Flow Diagrams (DFD). The integration and system testing stages of each program unit are integrated with each other and tested as a total system. Then the system will be sent to the system user. Website testing that has been designed uses testing with the Blackbox technique. The conclusion is that it can facilitate the patient quickly and efficiently. It is suggested that this system can be developed for the future so that it can be done electronically and the system can cover not only old patients, but also new patients so that it can be more effective and efficient in reducing long queues during registration at the hospital.

Dingo optimization algorithm-based random forests model to evaluate the compressive strength of the concrete at elevated temperatures

The qualities of the materials employed to manufacture concrete are significantly impacted by high temperatures, which results in a noticeable decrease in the material’s strength characteristics. Concrete must be worked very hard and allowed to reach the required compressive strength (fc). Nevertheless, a preliminary estimation of the desired outcome may be made with an outstanding degree of reliability by using supervised machine learning algorithms. The study combined the Dingo optimization algorithm (DOA), Coot bird optimization (COA), and Artificial rabbit optimization (ARO) with Random Forests (RF) evaluation to determine the fc of concrete at high temperatures. The abbreviations used for the combined methods are RFD, RFC, and RFA, respectively. Remarkably, removing the temperature (T) parameter from the input set leads to a remarkable 1100% improvement in the effectiveness index (PI) and normalized root mean squared error (NRMSE), while causing a significant fall in the coefficient of determination (R2). The findings suggest that all RFD, RFC, and RFA have substantial promise in properly forecasting the fc of concrete at high temperatures. More precisely, the RFD algorithm demonstrated exceptional precision with R2 values of 0.9885 and 0.9873 throughout the training and testing stages, respectively. Through a comparison of the error percentages for RFD, RFC, and RFA in error-based measurements, it becomes evident that RFD exhibits an error rate that is about 50% smaller compared to that of RFC and RFA. This prediction is crucial for various industries and applications where concrete structures are subjected to elevated temperatures, such as in fire resistance assessments for buildings, tunnels, bridges, and other infrastructure. By accurately forecasting the compressive strength of concrete under these conditions, engineers and designers can make informed decisions regarding the material’s suitability and performance in high-temperature environments, leading to enhanced safety, durability, and cost-effectiveness of structures.

A Dialogues Summarization Algorithm Based on Multi-task Learning

With the continuous advancement of social information, the number of texts in the form of dialogue between individuals has exponentially increased. However, it is very challenging to review the previous dialogue content before initiating a new conversation. In view of the above background, a new dialogue summarization algorithm based on multi-task learning is first proposed in the paper. Specifically, Minimum Risk Training is used as the loss function to alleviate the problem of inconsistent goals between the training phase and the testing phase. Then, in order to deal with the problem that the model cannot effectively distinguish gender pronouns, a gender pronoun discrimination auxiliary task based on contrast learning is designed to help the model learn to distinguish different gender pronouns. Finally, an auxiliary task of reducing exposure bias is introduced, which involves incorporating the summary generated during inference into another round of training to reduce the difference between the decoder inputs during the training and testing stages. Experimental results show that our model outperforms strong baselines on three public dialogue summarization datasets: SAMSUM, DialogSum, and CSDS.

Dual–ion symmetric hybrid supercapacitor built by vanadium oxide with expanded interlayers: The intercalation of cation–anion pair and reversible redox of anion excited by Al3+

The development of aluminum ion batteries (AIBs) is retarded by the low capacity of intercalation cathode and ineffective aluminum dissolution/deposition at Al anode. To solve the issues, benzoquinone(BQ)–preinserted hydrated vanadium oxide formulated as (BQ)0.04V2O5·0.87H2O (BQ-VO) with a large interlayer spacing of 13.3 Å was synthesized, which was used as both cathode and anode with Al(ClO4)3 acetonitrile (AN) electrolyte to fabricate a symmetric aluminum–ion supercapacitor. BQ-VO delivers a large capacity of ~152 mAh g−1 at 0.2 A g−1 in the range of 0–1.5 V, which is associated with the intercalation of ion (such as ClO4−, Al3+ or Al3+ − ClO4− pair) into BQ-VO and the reversible redox reaction of ClO4− ↔ Cl−, giving rise to a novel symmetric dual–ion hybrid supercapacitor. The reversible conversion of ClO4− ↔ Cl− is not observed in NaClO4 AN electrolyte, which is only excited by Al3+, leading to a continuous increasing capacity in the early stage of the cycling test. On the contrary, in the NaClO4 AN electrolyte, BQ-VO illustrates a surface-controlled capacitive performance with excellent cycling stability. Molecular dynamic (MD) simulations reveal that Al3+ and Na+ possess different solvation/anion sheaths, leading to different electrochemical performances of BQ-VO in the Al(ClO4)3 and NaClO4 electrolytes. The present work provides a routine to utilize the redox reaction of halide for energy storage.

Training of a Neural Network System in the Task of Detecting Blue Stains in a Sawmill Wood Inspection System

This article presents the operation of an automatic pine sawn timber inspection system, which was developed at the Woodinspector company and is offered commercially. The vision inspection system is used to detect various wood defects, including knots, blue stain, and mechanical damage caused by worms. A blue stain is a defect that is difficult to detect based on the color of the wood, because it can be easily confused with wood defects or dirt that do not impair its strength properties. In particular, the issues of detecting blue stain in wood, the use of artificial neural networks, and improving the operation of the system in production conditions are discussed in this article. While training the network, 400 boards, 4 m long, and their cross-sections of 100 × 25 [mm] were used and photographed using special scanners with laser illuminators from four sides. The test stages were carried out during an 8-hour workday at a sawmill (8224 m of material was scanned) on material with an average of 10% blue stain (every 10th board has more than 30% of its length stained blue). The final learning error was assessed based on defective boards detected by humans after the automatic selection stage. The system error for 5387 boards, 550 m long, which had blue staining that was not detected by the scanner (clean) was 0.4% (25 pieces from 5387), and 0.1 % in the case of 3412 boards, 610 mm long, on which there were no blue stains, but were wrongly classified (blue stain). For 6491 finger-joint boards (180–400 mm), 48 pieces were classified as class 1 (clean), but had a blue stain (48/6491 = 0.7%), and 28 pieces did not have a blue stain, but were classified as class 2 (28/3561 = 0.7%).

Improving span-based Aspect Sentiment Triplet Extraction with part-of-speech filtering and contrastive learning

Aspect Sentiment Triple Extraction (ASTE), a subtask of fine-grained sentiment analysis, aims to extract aspect terms, opinion terms, and their corresponding sentiment polarities from sentences. Previous methods often enumerated all possible spans of aspects and opinions as candidate spans that contain many invalid and irrelevant spans. This made the model training and prediction more difficult due to noised spans, leading to poor performance. To address this issue, we first propose a novel span-level approach that explicitly considers prior grammatical knowledge to generate possible candidate spans by part-of-speech filtering. In this way, our approach can make the model easier to be trained and achieve higher performance at the test stage. Besides, the quality of span-level representation of aspects and opinions is crucial for predicting their sentiment relation. To build a high-quality span-level representation of aspects and opinions, we first incorporate the contextual embedding of the entire sequence into span-level representations. Then, we introduce an auxiliary loss based on contrastive learning to make a more compact representation of the same polarities. Experimental evaluations on the 14Lap, 14Res, 15Res, and 16Res datasets demonstrate the effectiveness of our model, achieving state-of-the-art performance in span-based triplet extraction.