training model
Recently Published Documents


TOTAL DOCUMENTS

2625
(FIVE YEARS 1100)

H-INDEX

31
(FIVE YEARS 7)

2022 ◽  
Vol 40 (3) ◽  
pp. 1-29
Author(s):  
Jing Yao ◽  
Zhicheng Dou ◽  
Ji-Rong Wen

Personalized search tailors document ranking lists for each individual user based on her interests and query intent to better satisfy the user’s information need. Many personalized search models have been proposed. They first build a user interest profile from the user’s search history, and then re-rank the documents based on the personalized matching scores between the created profile and candidate documents. In this article, we attempt to solve the personalized search problem from an alternative perspective of clarifying the user’s intention of the current query. We know that there are many ambiguous words in natural language such as “Apple.” People with different knowledge backgrounds and interests have personalized understandings of these words. Therefore, we propose a personalized search model with personal word embeddings for each individual user that mainly contain the word meanings that the user already knows and can reflect the user interests. To learn great personal word embeddings, we design a pre-training model that captures both the textual information of the query log and the information about user interests contained in the click-through data represented as a graph structure. With personal word embeddings, we obtain the personalized word and context-aware representations of the query and documents. Furthermore, we also employ the current session as the short-term search context to dynamically disambiguate the current query. Finally, we use a matching model to calculate the matching score between the personalized query and document representations for ranking. Experimental results on two large-scale query logs show that our designed model significantly outperforms state-of-the-art personalization models.


10.29007/h46n ◽  
2022 ◽  
Author(s):  
Hoang Nhut Huynh ◽  
Minh Thanh Do ◽  
Gia Thinh Huynh ◽  
Anh Tu Tran ◽  
Trung Nghia Tran

Diabetic retinopathy (DR) is a complication of diabetes mellitus that causes retinal damage that can lead to vision loss if not detected and treated promptly. The common diagnosis stages of the disease take time, effort, and cost and can be misdiagnosed. In the recent period with the explosion of artificial intelligence, deep learning has become the most popular tool with high performance in many fields, especially in the analysis and classification of medical images. The Convolutional Neural Network (CNN) is more widely used as a deep learning method in medical imaging analysis with highly effective. In this paper, the five-stage image of modern DR (healthy, mild, moderate, severe, and proliferative) can be detected and classified using the deep learning technique. After cross-validation training and testing on the corresponding 5,590-image dataset, a pre-MobileNetV2 training model is proposed in classifying stages of diabetic retinopathy. The average accuracy of the model achieved was 93.89% with the precision of 94.00%, recall 92.00% and f1-score 90.00%. The corresponding thermal image is also given to help experts for evaluating the influence of the retina in each different stage.


2022 ◽  
Vol 6 (1) ◽  
pp. 47-51
Author(s):  
Dawn Fei-yue Tsang

This research project aims at consolidating and revitalizing translation and interpreting pedagogy with dual emphasis on the cultural soft power and international discourse rights of China via emphasizing international competitiveness. In order to realize the significant needs of the current market requirements and the new era of China, this project proposes the “ICC” interpreting training model. It refers to the combination of intercultural communication competency with international competitiveness as the teaching and learning outcomes by means of integrating the following five components in the course content and the whole curriculum design: (1) interpreting competency – bilingual competence and interpreting skill-based training; e.g., short-term memory and note-taking; (2) national value – strengthening translation and the crucial roles and responsibilities of interpreter trainers and trainees in disseminating national culture; (3) expertise of subject matters – equipping students with expertise for work field and meeting the market requirements; e.g., specific professional knowledge and jargons demanded by the tasks; (4) professional ethics and image – extra-linguistic knowledge emphasizing a translator’s and an interpreter’s professional code of conduct and ethics in a way that can establish the “iconic” image of a professional interpreter; e.g., confidentiality, posture, and appearance; (5) practical assessment – evaluation of students’ performance in practical translation and interpreting opportunities and/or internship in organizations of various natures according to international standards. This research proposes a new training model to incorporate intercultural communication competency with international competitiveness. The significant role of interpreting pedagogy in contributing to a sound national name is investigated. The new “ICC” model that this research is proposing answers such a call for the significant role of raising cultural soft power and international discourse rights in China.  


Sensors ◽  
2022 ◽  
Vol 22 (2) ◽  
pp. 545
Author(s):  
Bor-Jiunn Hwang ◽  
Hui-Hui Chen ◽  
Chaur-Heh Hsieh ◽  
Deng-Yu Huang

Based on experimental observations, there is a correlation between time and consecutive gaze positions in visual behaviors. Previous studies on gaze point estimation usually use images as the input for model trainings without taking into account the sequence relationship between image data. In addition to the spatial features, the temporal features are considered to improve the accuracy in this paper by using videos instead of images as the input data. To be able to capture spatial and temporal features at the same time, the convolutional neural network (CNN) and long short-term memory (LSTM) network are introduced to build a training model. In this way, CNN is used to extract the spatial features, and LSTM correlates temporal features. This paper presents a CNN Concatenating LSTM network (CCLN) that concatenates spatial and temporal features to improve the performance of gaze estimation in the case of time-series videos as the input training data. In addition, the proposed model can be optimized by exploring the numbers of LSTM layers, the influence of batch normalization (BN) and global average pooling layer (GAP) on CCLN. It is generally believed that larger amounts of training data will lead to better models. To provide data for training and prediction, we propose a method for constructing datasets of video for gaze point estimation. The issues are studied, including the effectiveness of different commonly used general models and the impact of transfer learning. Through exhaustive evaluation, it has been proved that the proposed method achieves a better prediction accuracy than the existing CNN-based methods. Finally, 93.1% of the best model and 92.6% of the general model MobileNet are obtained.


Sensors ◽  
2022 ◽  
Vol 22 (2) ◽  
pp. 537
Author(s):  
Caiyue Zhou ◽  
Yanfen Kong ◽  
Chuanyong Zhang ◽  
Lin Sun ◽  
Dongmei Wu ◽  
...  

Group-based sparse representation (GSR) uses image nonlocal self-similarity (NSS) prior to grouping similar image patches, and then performs sparse representation. However, the traditional GSR model restores the image by training degraded images, which leads to the inevitable over-fitting of the data in the training model, resulting in poor image restoration results. In this paper, we propose a new hybrid sparse representation model (HSR) for image restoration. The proposed HSR model is improved in two aspects. On the one hand, the proposed HSR model exploits the NSS priors of both degraded images and external image datasets, making the model complementary in feature space and the plane. On the other hand, we introduce a joint sparse representation model to make better use of local sparsity and NSS characteristics of the images. This joint model integrates the patch-based sparse representation (PSR) model and GSR model, while retaining the advantages of the GSR model and the PSR model, so that the sparse representation model is unified. Extensive experimental results show that the proposed hybrid model outperforms several existing image recovery algorithms in both objective and subjective evaluations.


2022 ◽  
Vol 2022 ◽  
pp. 1-11
Author(s):  
Rui Yang ◽  
Zenghui An ◽  
Shijun Song

A convolutional neural network has the characteristics of sharing information between layers, which can realize high-dimensional data processing. In general, the convolutional neural network uses a feedback mechanism to realize parameter self-regulation, which solves the disadvantages of manual parameter adjustment. However, it is unable to determine the iteration number with the best calculation accuracy. Calculation efficiency cannot be guaranteed while achieving the best accuracy. In this paper, a multilayer extreme learning convolutional neural network model is proposed for feature recognition and classification. Firstly, two-dimensional spatial characteristics of planetary bearing status data were enhanced. Then, extreme learning machine is embedded in a convolution layer to solve convex optimization problems. Finally, the parameters obtained from the training model were nested into a network to initialize the model parameters to separate each status feature. Planetary bearing experimental cases show the effectiveness and superiority of the proposed model in the recognition and classification of weak signals.


2022 ◽  
pp. 155335062110689
Author(s):  
Shotaro Okachi ◽  
Takayasu Ito ◽  
Kazuhide Sato ◽  
Shingo Iwano ◽  
Yuka Shinohara ◽  
...  

Background/need. The increases in reference images and information during bronchoscopy using virtual bronchoscopic navigation (VBN) and fluoroscopy has potentially created the need for support using a head-mounted display (HMD) because bronchoscopists feel difficulty to see displays that are at a distance from them and turn their head and body in various directions. Methodology and device description. The binocular see-through Moverio BT-35E Smart Glasses can be connected via a high-definition multimedia interface and have a 720p high-definition display. We developed a system that converts fluoroscopic (live and reference), VBN, and bronchoscopic image signals through a converter and references them using the Moverio BT-35E. Preliminary results. We performed a virtual bronchoscopy-guided transbronchial biopsy simulation using the system. Four experienced pulmonologists performed a simulated bronchoscopy of 5 cases each with the Moverio BT-35E glasses, using bronchoscopy training model. For all procedures, the bronchoscope was advanced successfully into the target bronchus according to the VBN image. None of the operators reported eye or body fatigue during or after the procedure. Current status. This small-scale simulation study suggests the feasibility of using a HMD during bronchoscopy. For clinical use, it is necessary to evaluate the safety and usefulness of the system in larger clinical trials in the future.


Autism ◽  
2022 ◽  
pp. 136236132110655
Author(s):  
Sarah R Rieth ◽  
Kelsey S Dickson ◽  
Jordan Ko ◽  
Rachel Haine-Schlagel ◽  
Kim Gaines ◽  
...  

Best-practice recommendations for young children at high likelihood of autism include active involvement of caregivers in intervention. However, the use of evidence-based parent-mediated interventions in community practice remains limited. Preliminary evidence suggests that Project ImPACT for Toddlers demonstrates positive parent and child outcomes in community settings. Project ImPACT for Toddlers was adapted specifically for toddlers and teaches parents of young children strategies to build their child’s social, communication, and play skills in daily routines. This study reports implementation outcomes from the initial community rollout of Project ImPACT for Toddlers and examines the system-wide intervention reach, with the goal of informing continued community sustainment and scale-up. Participants include 38 community providers who participated in a Project ImPACT for Toddlers’ training study who completed an implementation survey and semi-structured interviews after approximately 3 months of community implementation. Participants perceived the training model as acceptable and appropriate, and identified several strengths of the approach. Interview themes also supported the feasibility, acceptability, and utility of the intervention in community settings. Quantitative findings complemented the thematic results from interviews. Intervention reach data indicate an increasing number of agencies delivering and families receiving Project ImPACT for Toddlers. Efforts to scale-up evidence-based interventions in early intervention should continue to build upon the model of the Bond, Regulate, Interact, Develop, Guide, and Engage Collaborative. Lay abstract Expert recommendations for toddlers who are likely to develop autism include caregivers being actively involved in the services children receive. However, many services available in the community may not follow these recommendations. Evidence suggests that an intervention named Project ImPACT for Toddlers demonstrates positive parent and child outcomes for families in the community. Project ImPACT for Toddlers was designed specifically for toddlers by a group of parents, clinicians, researchers, and funders. It teaches parents of young children strategies to support their child’s development in daily routines. This study reports the perspectives of early intervention providers who learned to use Project ImPACT for Toddlers on whether the intervention was a good fit for their practice and easy to use. The study also examines how many agencies are using Project ImPACT for Toddlers and how many families have received the intervention in the community. The goal of the study is to inform the continued use of Project ImPACT for Toddlers in the community and support offering the intervention in other regions. Participants include 38 community providers who participated in a training study of Project ImPACT for Toddlers and completed a survey and semi-structured interview after approximately 3 months of using Project ImPACT for Toddlers with families. Participants perceived the training model as acceptable and appropriate, and identified the group-based model of training, comprehensive materials, and agency support as strengths of the approach. Survey findings complemented the results from the interviews. Data indicate an increasing number of agencies and families accessing Project ImPACT for Toddlers. Efforts to expand evidence-based intervention in early intervention should continue to build upon the model used for Project ImPACT for Toddlers.


2022 ◽  
Vol 2022 ◽  
pp. 1-9
Author(s):  
Jiahui Gu

The traditional mixed oral English teaching model has many obvious shortcomings, such as the inability to correct the students’ oral pronunciation errors and feed them back in time, which leads to the slow improvement of students’ English learning level. For this reason, this paper proposes a guided teaching model based on core literacy. According to the structure of the oral English mixed teaching model, determine the application plan of the oral English mixed teaching model, design the development environment, obtain the corpus, design the oral training model, extract the oral features, identify the wrong pronunciation and correct it in time, clarify the evaluation purpose, obtain preliminary evaluation indicators, reduce evaluation indicators and determine indicator weights, obtain indicator feature information, generate fuzzy rules, obtain fuzzy matrices, achieve quantitative evaluation, and synthesize all evaluation scores to construct a result vector matrix to realize the study of mixed spoken language teaching mode. Research shows that the mixed teaching method is effective and feasible and can effectively improve the accuracy of the evaluation results of the mixed oral English teaching model.


2022 ◽  
Author(s):  
FAZIL APAYDIN ◽  
Meshari Saghir ◽  
Rodrigo Fortunato Fernandez Pellon Garcia ◽  
Mahmoud Daoud ◽  
Ayman Jaber

Abstract: Background: Septoplasty and rhinoplasty are difficult operations to learn and teach. Many modalities have been proposed to make the teaching process of these operations easier. In this study, it was investigated if lamb heads were good training models to teach septoplasty and rhinoplasty to trainees or experienced surgeons. Methods: In the first part of the study, 21 lamb heads were dissected according to a dissection protocol and several anatomical distances were measured in order to compare them with human cadavers. In the second project 8 lamb heads were dissected and different preservation rhinoplasty techniques were practiced. Results: The study on 21 lamb heads used showed that the lateral crura were 17.8 x 11.6, average interdomal distance was 8.1 mm, average domal width was 3.7 mm. The average length of the upper lateral cartilages was 31.1 mm laterally and 21.2 medially. The average length of the nasal bones was 63.9 mm and the width was 16 mm. In the second part of the study 8 lamb heads were used to experience where high strip techniques were used in 5 and Cottle technique in 3. Conclusion: This study revealed that lamb head should be considered as an excellent training model for septoplasty and rhinoplasty. Its very low cost, ease of availability, and close similarity to the human cadavers can be counted as the main advantages. This study also proved that it was not only a tool for beginners, but also a very helpful tool for experienced surgeons to try new methods.


Sign in / Sign up

Export Citation Format

Share Document