learning ability
Recently Published Documents


TOTAL DOCUMENTS

2617
(FIVE YEARS 1071)

H-INDEX

61
(FIVE YEARS 8)

2022 ◽  
Vol 23 (2) ◽  
pp. 972
Author(s):  
Chen Jin ◽  
Zhuangwei Shi ◽  
Chuanze Kang ◽  
Ken Lin ◽  
Han Zhang

X-ray diffraction technique is one of the most common methods of ascertaining protein structures, yet only 2–10% of proteins can produce diffraction-quality crystals. Several computational methods have been proposed so far to predict protein crystallization. Nevertheless, the current state-of-the-art computational methods are limited by the scarcity of experimental data. Thus, the prediction accuracy of existing models hasn’t reached the ideal level. To address the problems above, we propose a novel transfer-learning-based framework for protein crystallization prediction, named TLCrys. The framework proceeds in two steps: pre-training and fine-tuning. The pre-training step adopts attention mechanism to extract both global and local information of the protein sequences. The representation learned from the pre-training step is regarded as knowledge to be transferred and fine-tuned to enhance the performance of crystalization prediction. During pre-training, TLCrys adopts a multi-task learning method, which not only improves the learning ability of protein encoding, but also enhances the robustness and generalization of protein representation. The multi-head self-attention layer guarantees that different levels of the protein representation can be extracted by the fine-tuned step. During transfer learning, the fine-tuning strategy used by TLCrys improves the task-specialized learning ability of the network. Our method outperforms all previous predictors significantly in five crystallization stages of prediction. Furthermore, the proposed methodology can be well generalized to other protein sequence classification tasks.


2022 ◽  
Vol 2 (2) ◽  
pp. 83-89
Author(s):  
Partonduhan aritonang Partonduhan aritonang ◽  
Parsaoran Tamba ◽  
Jemmy Charles Kewas

PENGARUH GAME ONLINE TERHADAP CARA BELAJAR MAHASISWA JURUSAN PENDIDIKAN TEKNIK MESIN UNIVERSITAS NEGERI MANADO Partonduhan Aritonang1, I. P. Tamba2, Jemmy Charles Kelas3 1,2,3Jurusan Pendidikan Teknik Mesin, Universitas Negeri Manado, Kab. Minahasa e-mail: [email protected], [email protected], [email protected]   ABSTRAK Mahasiswa Pendidikan Teknik Mesin Universitas Negeri Manado yang merupakan anak-anak perantau kini telah mendapatkan dampak yang sangat nyata dari permainan game online. Terbukti dari banyaknya mahasiswa yang ikut ambil bagian dalam permainan ini, dari hasil pengamatan peneliti selaku mahasiswa yang aktif mendapatkan banyak data bahwa mahasiswa Pendidikan Teknik Mesin Universitas Negeri Manado yang aktif bermain memiliki kemampuan cara belajar yang kurang aktif dalam pembelajaran. Penelitian ini menggunakan metode penelitian deskriptif kuantitatif. Metode pengumpulan data yang digunakan yakni kuisioner atau angket. Teknik analisi data yang digunakan dalam penelitian ini yaitu analisis statistik deskriptif, Teknik Analisis Regresi dan pengujian hipotesis.Hasil dari penelitian ini yakni : bahwa pengaruh game online (X) terhadap cara belajar mahasiswa (Y) pada taraf t hitung > t tabel dan hasil uji korelasi rxy 0849. Game online berpengaruh signifikan terhadap cara belajar. Ini dapat dibuktikan dari hasil nilai Fhitung sebesar 4.113 dan nilai signifikansi Ftabel 0.00 < 0.05. Besarnya koefisien determinasi sebesar 0.79 atau 79%. Hal ini berarti 79% pengaruh game online terhadap cara belajar mahasiswa sedangkan untuk selebihnya 21% dipengaruhi oleh variabel lain yang tidak diteliti oleh penelitian ini.   Kata kunci : Game Online, Cara Belajar Mahasiswa THE INFLUENCE OF ONLINE GAMES ON HOW STUDENTS STUDYING MECHANICAL ENGINEERING AT MANADO STATE UNIVERSITY ABSTRACT Manado university's advanced mechanical engineering student who is a migrant child has now had a very real impact on online gaming. It is evident from the many students participating in the game that researchers as active university students have received a wealth of data that students studying engineering at manado state university who actively play have a learning ability that is less active in learning. The study USES a quantitative descriptive study method. The data collection method used was "questionnaire or angket." The data analysis used in the study are descriptive statistical analysis, regression analysis and hypothetical testing. The results of this study are: that how online games affect students' learning (y) at a level of t count > t tables and rxy 0849 cordating results. Online games significantly affect how to learn. This can be verified from the results of the ftable value of 4,113 and the significance of ftable 0.00. Critical coefficiencies by 0.79 or 79%. This means 79% of the impact online games have on student learning while for the rest 21% are affected by other variables not examined by this study. Key words : Game Online, student learning


2022 ◽  
Author(s):  
Yufan Zhang ◽  
Honglin Wen ◽  
Qiuwei Wu ◽  
Qian Ai

Prediction intervals (PIs) offer an effective tool for quantifying uncertainty of loads in distribution systems. The traditional central PIs cannot adapt well to skewed distributions, and their offline training fashion is vulnerable to the unforeseen change in future load patterns. Therefore, we propose an optimal PI estimation approach, which is online and adaptive to different data distributions by adaptively determining symmetric or asymmetric probability proportion pairs for quantiles of PIs’ bounds. It relies on the online learning ability of reinforcement learning (RL) to integrate the two online tasks, i.e., the adaptive selection of probability proportion pairs and quantile predictions, both of which are modeled by neural networks. As such, the quality of quantiles-formed PI can guide the selection process of optimal probability proportion pairs, which forms a closed loop to improve PIs’ quality. Furthermore, to improve the learning efficiency of quantile forecasts, a prioritized experience replay (PER) strategy is proposed for online quantile regression processes. Case studies on both load and net load demonstrate that the proposed method can better adapt to data distribution compared with online central PIs method. Compared with offline-trained methods, it obtains PIs with better quality and is more robust against concept drift.


2022 ◽  
Author(s):  
Yufan Zhang ◽  
Honglin Wen ◽  
Qiuwei Wu ◽  
Qian Ai

Prediction intervals (PIs) offer an effective tool for quantifying uncertainty of loads in distribution systems. The traditional central PIs cannot adapt well to skewed distributions, and their offline training fashion is vulnerable to the unforeseen change in future load patterns. Therefore, we propose an optimal PI estimation approach, which is online and adaptive to different data distributions by adaptively determining symmetric or asymmetric probability proportion pairs for quantiles of PIs’ bounds. It relies on the online learning ability of reinforcement learning (RL) to integrate the two online tasks, i.e., the adaptive selection of probability proportion pairs and quantile predictions, both of which are modeled by neural networks. As such, the quality of quantiles-formed PI can guide the selection process of optimal probability proportion pairs, which forms a closed loop to improve PIs’ quality. Furthermore, to improve the learning efficiency of quantile forecasts, a prioritized experience replay (PER) strategy is proposed for online quantile regression processes. Case studies on both load and net load demonstrate that the proposed method can better adapt to data distribution compared with online central PIs method. Compared with offline-trained methods, it obtains PIs with better quality and is more robust against concept drift.


2022 ◽  
Vol 1 ◽  
Author(s):  
Junchao Lei ◽  
Tao Lei ◽  
Weiqiang Zhao ◽  
Mingyuan Xue ◽  
Xiaogang Du ◽  
...  

Deep convolutional neural networks (DCNNs) have been widely used in medical image segmentation due to their excellent feature learning ability. In these DCNNs, the pooling operation is usually used for image down-sampling, which can gradually reduce the image resolution and thus expands the receptive field of convolution kernel. Although the pooling operation has the above advantages, it inevitably causes information loss during the down-sampling of the pooling process. This paper proposes an effective weighted pooling operation to address the problem of information loss. First, we set up a pooling window with learnable parameters, and then update these parameters during the training process. Secondly, we use weighted pooling to improve the full-scale skip connection and enhance the multi-scale feature fusion. We evaluated weighted pooling on two public benchmark datasets, the LiTS2017 and the CHAOS. The experimental results show that the proposed weighted pooling operation effectively improve network performance and improve the accuracy of liver and liver-tumor segmentation.


2022 ◽  
Vol 12 ◽  
Author(s):  
Minghui Du ◽  
Yiqun Qian

The study aims to explore the roles of Massive Open Online Courses (MOOCs) based on deep learning in college students’ English grammar teaching. The data are collected using a survey. After the experimental data are analyzed, it is found that students have a low sense of happiness and satisfaction and are unwilling to practice oral English and learn language points in English learning. They think that college English learning only meets the needs of CET-4 and CET-6 and does not take it as the ultimate learning goal. After the necessity and problems in English grammar teaching are discussed, the advantages of flipped classrooms of MOOCs are discussed in English grammar teaching. A teaching platform is constructed to study the foreign language teaching mode under MOOCs, and classroom teaching is combined with the advantages of MOOCs following the principle of “teaching students according to their personalities” to improve the listening, speaking, reading, writing, and translation skills of foreign language majors. The results show that high-quality online teaching resources and the deep learning-based teaching environment can provide a variety of interactive tools, by which students can communicate with their peers and teachers online. Sharing open online communication, classroom discussion, and situational simulation can enhance teachers’ deep learning ability, like the ability to communication and transfer thoughts. Constructivism with interaction as the core can help students grasp new knowledge easily. Extensive communication and interaction are important ways for learning and thinking. The new model provides students with profound learning experience, expands the teaching resources of MOOCs around the world, and maximizes the interaction between online and offline teachers and students, making knowledge widely rooted in the campus and realizing the combination of online resources and campus classroom teaching. Students can learn the knowledge through autonomous learning and discussion before class, which greatly broadens the learning time and space. In the classroom and after class, the internalization and sublimation of knowledge are completed through group cooperation, inquiry learning, scenario simulation, display, and evaluation, promoting students to know about new knowledge and highlighting the dominant position of students.


2022 ◽  
Vol 9 ◽  
Author(s):  
Huan-Yu Liu ◽  
Juanjuan Guo ◽  
Chang Zeng ◽  
Yuming Cao ◽  
Ruoxi Ran ◽  
...  

Background: Long-term effects of Coronavirus Disease 2019 (COVID-19) on infants born to infected mothers are not clear. Fine motor skills are crucial for the development of infant emotional regulation, learning ability and social skills.Methods: Clinical information of 100 infants born to 98 mothers (COVID-19 n = 31, non-COVID-19 n = 67) were collected. Infants were follow-up up to 9 months post-partum. The placental tissues were examined for SARS-CoV-2 infection, pathological changes, cytokines, and mtDNA content.Results: Decreased placental oxygen and nutrient transport capacity were found in infected pregnant women. Increased IL-2, IL-6, TNF-α, and IFN-γ were detected in trophoblast cells and maternal blood of COVID-19 placentas. Elevated early fine motor abnormal-ities and increased serum TNI (troponin I) levels at delivery were observed in infants born to mothers with COVID-19. Increased abnormal mitochondria and elevated mtDNA content were found in the placentas from infected mothers. The placental mtDNA content of three infants with abnormal DDST were increased by 4, 7, and 10%, respectively, compared to the mean of the COVID-19 group. The Maternal Vascular Malperfusion (MVM), elevated cytokines and increased placental mtDNA content in mothers with COVID-19 might be associated with transient early fine motor abnormalities in infants. These abnormalities are only temporary, and they could be corrected by daily training.Conclusions: Babies born to COVID-19 mothers with mild symptoms appeared to have little or no excess long-term risks of abnormal physical and neurobehavioral development as compared with the infants delivered by non-COVID-19 mothers.


2022 ◽  
Author(s):  
Chuan Xu ◽  
Jian Gao ◽  
Jiaxin Gao ◽  
Lingling Li ◽  
Fangping He ◽  
...  

When listening to an unknown language, listeners could learn the transitional probability between syllables and group frequently co-occurred syllables into a whole unit. Such statistical learning ability has been demonstrated for both pre-verbal infants and adults, even during passive listening. Here, we investigated whether statistical learning occurred in patients in minimally conscious state (MCS) and patients emerged from the minimally conscious state (EMCS) using electroencephalography (EEG). We presented to participants an isochronous sequence of syllables, which were composed of either 2-word real phrases or 2-word artificial phrases that were defined by the transitional probability between words. An inter-trial phase coherence (ITPC) analysis revealed that the phrase-rate EEG response was weakened in EMCS patients compared with healthy individuals, and was even more severely weakened in MCS patients. Although weak, the phrase-rate response or its harmonics remained statistically significant in MCS patients, suggesting that the statistical learning ability was preserved in MCS patients. The word-rate response was also weakened with a decreased level of consciousness. The harmonics of the word-rate response, however,were more salient in MCS than EMCS patients in the alpha and beta bands. Together with previous studies, the current results suggest that MCS patients retain residual learning ability, which can potentially be harnessed to induce neural plasticity, and that different frequency bands are differentially related to the consciousness level.


2022 ◽  
Vol 2022 ◽  
pp. 1-9
Author(s):  
Shaobin Ma ◽  
Lan Li ◽  
Chengwen Zhang

Effective noise removal has become a hot topic in image denoising research while preserving important details of an image. An adaptive threshold image denoising algorithm based on fitting diffusion is proposed. Firstly, the diffusion coefficient in the diffusion equation is improved, and the fitting diffusion coefficient is established to overcome the defects of texture detail loss and edge degradation caused by excessive diffusion intensity. Then, the threshold function is adaptively designed and improved so that it can automatically control the threshold of the function according to the maximum gray value of the image and the number of iterations, so as to further preserve the important details of the image such as edge and texture. A neural network is used to realize image denoising because of its good learning ability of image statistical characteristics, mainly by the diffusion equation and deep learning (CNN) algorithm as the foundation, focus on the effects of activation function of network optimization, using multiple feature extraction technology in-depth networks to study the characteristics of the input image richer, and how to better use the adaptive algorithm on the depth of diffusion equation and optimization backpropagation learning. The training speed of the model is accelerated and the convergence of the algorithm is improved. Combined with batch standardization and residual learning technology, the image denoising network model based on deep residual learning of the convolutional network is designed with better denoising performance. Finally, the algorithm is compared with other excellent denoising algorithms. From the comparison results, it can be seen that the improved denoising algorithm in this paper can also improve the detail restoration of denoised images without losing the sharpness. Moreover, it has better PSNR than other excellent denoising algorithms at different noise standard deviations. The PSNR of the new algorithm is greatly improved compared with the classical algorithm, which can effectively suppress the noise and protect the image edge and detail information.


Sign in / Sign up

Export Citation Format

Share Document