A Proof of Badd Handwriting in MNIST Training Dataset Making CNN to Predict Goodd Handwriting Wrong

2017 ◽  
Author(s):  
Xinbin Zhang
Keyword(s):  
2020 ◽  
Vol 2020 (10) ◽  
pp. 181-1-181-7
Author(s):  
Takahiro Kudo ◽  
Takanori Fujisawa ◽  
Takuro Yamaguchi ◽  
Masaaki Ikehara

Image deconvolution has been an important issue recently. It has two kinds of approaches: non-blind and blind. Non-blind deconvolution is a classic problem of image deblurring, which assumes that the PSF is known and does not change universally in space. Recently, Convolutional Neural Network (CNN) has been used for non-blind deconvolution. Though CNNs can deal with complex changes for unknown images, some CNN-based conventional methods can only handle small PSFs and does not consider the use of large PSFs in the real world. In this paper we propose a non-blind deconvolution framework based on a CNN that can remove large scale ringing in a deblurred image. Our method has three key points. The first is that our network architecture is able to preserve both large and small features in the image. The second is that the training dataset is created to preserve the details. The third is that we extend the images to minimize the effects of large ringing on the image borders. In our experiments, we used three kinds of large PSFs and were able to observe high-precision results from our method both quantitatively and qualitatively.


2020 ◽  
Vol 27 ◽  
Author(s):  
Zaheer Ullah Khan ◽  
Dechang Pi

Background: S-sulfenylation (S-sulphenylation, or sulfenic acid) proteins, are special kinds of post-translation modification, which plays an important role in various physiological and pathological processes such as cytokine signaling, transcriptional regulation, and apoptosis. Despite these aforementioned significances, and by complementing existing wet methods, several computational models have been developed for sulfenylation cysteine sites prediction. However, the performance of these models was not satisfactory due to inefficient feature schemes, severe imbalance issues, and lack of an intelligent learning engine. Objective: In this study, our motivation is to establish a strong and novel computational predictor for discrimination of sulfenylation and non-sulfenylation sites. Methods: In this study, we report an innovative bioinformatics feature encoding tool, named DeepSSPred, in which, resulting encoded features is obtained via n-segmented hybrid feature, and then the resampling technique called synthetic minority oversampling was employed to cope with the severe imbalance issue between SC-sites (minority class) and non-SC sites (majority class). State of the art 2DConvolutional Neural Network was employed over rigorous 10-fold jackknife cross-validation technique for model validation and authentication. Results: Following the proposed framework, with a strong discrete presentation of feature space, machine learning engine, and unbiased presentation of the underline training data yielded into an excellent model that outperforms with all existing established studies. The proposed approach is 6% higher in terms of MCC from the first best. On an independent dataset, the existing first best study failed to provide sufficient details. The model obtained an increase of 7.5% in accuracy, 1.22% in Sn, 12.91% in Sp and 13.12% in MCC on the training data and12.13% of ACC, 27.25% in Sn, 2.25% in Sp, and 30.37% in MCC on an independent dataset in comparison with 2nd best method. These empirical analyses show the superlative performance of the proposed model over both training and Independent dataset in comparison with existing literature studies. Conclusion : In this research, we have developed a novel sequence-based automated predictor for SC-sites, called DeepSSPred. The empirical simulations outcomes with a training dataset and independent validation dataset have revealed the efficacy of the proposed theoretical model. The good performance of DeepSSPred is due to several reasons, such as novel discriminative feature encoding schemes, SMOTE technique, and careful construction of the prediction model through the tuned 2D-CNN classifier. We believe that our research work will provide a potential insight into a further prediction of S-sulfenylation characteristics and functionalities. Thus, we hope that our developed predictor will significantly helpful for large scale discrimination of unknown SC-sites in particular and designing new pharmaceutical drugs in general.


2019 ◽  
Vol 24 (34) ◽  
pp. 4013-4022 ◽  
Author(s):  
Xiang Cheng ◽  
Xuan Xiao ◽  
Kuo-Chen Chou

Knowledge of protein subcellular localization is vitally important for both basic research and drug development. With the avalanche of protein sequences emerging in the post-genomic age, it is highly desired to develop computational tools for timely and effectively identifying their subcellular localization based on the sequence information alone. Recently, a predictor called “pLoc-mPlant” was developed for identifying the subcellular localization of plant proteins. Its performance is overwhelmingly better than that of the other predictors for the same purpose, particularly in dealing with multi-label systems in which some proteins, called “multiplex proteins”, may simultaneously occur in two or more subcellular locations. Although it is indeed a very powerful predictor, more efforts are definitely needed to further improve it. This is because pLoc-mPlant was trained by an extremely skewed dataset in which some subsets (i.e., the protein numbers for some subcellular locations) were more than 10 times larger than the others. Accordingly, it cannot avoid the biased consequence caused by such an uneven training dataset. To overcome such biased consequence, we have developed a new and bias-free predictor called pLoc_bal-mPlant by balancing the training dataset. Cross-validation tests on exactly the same experimentconfirmed dataset have indicated that the proposed new predictor is remarkably superior to pLoc-mPlant, the existing state-of-the-art predictor in identifying the subcellular localization of plant proteins. To maximize the convenience for the majority of experimental scientists, a user-friendly web-server for the new predictor has been established at http://www.jci-bioinfo.cn/pLoc_bal-mPlant/, by which users can easily get their desired results without the need to go through the detailed mathematics.


2019 ◽  
Vol 17 ◽  
Author(s):  
Yanqiu Yao ◽  
Xiaosa Zhao ◽  
Qiao Ning ◽  
Junping Zhou

Background: Glycation is a nonenzymatic post-translational modification process by attaching a sugar molecule to a protein or lipid molecule. It may impair the function and change the characteristic of the proteins which may lead to some metabolic diseases. In order to understand the underlying molecular mechanisms of glycation, computational prediction methods have been developed because of their convenience and high speed. However, a more effective computational tool is still a challenging task in computational biology. Methods: In this study, we showed an accurate identification tool named ABC-Gly for predicting lysine glycation sites. At first, we utilized three informative features, including position-specific amino acid propensity, secondary structure and the composition of k-spaced amino acid pairs to encode the peptides. Moreover, to sufficiently exploit discriminative features thus can improve the prediction and generalization ability of the model, we developed a two-step feature selection, which combined the Fisher score and an improved binary artificial bee colony algorithm based on support vector machine. Finally, based on the optimal feature subset, we constructed the effective model by using Support Vector Machine on the training dataset. Results: The performance of the proposed predictor ABC-Gly was measured with the sensitivity of 76.43%, the specificity of 91.10%, the balanced accuracy of 83.76%, the area under the receiver-operating characteristic curve (AUC) of 0.9313, a Matthew’s Correlation Coefficient (MCC) of 0.6861 by 10-fold cross-validation on training dataset, and a balanced accuracy of 59.05% on independent dataset. Compared to the state-of-the-art predictors on the training dataset, the proposed predictor achieved significant improvement in the AUC of 0.156 and MCC of 0.336. Conclusion: The detailed analysis results indicated that our predictor may serve as a powerful complementary tool to other existing methods for predicting protein lysine glycation. The source code and datasets of the ABC-Gly were provided in the Supplementary File 1.


2019 ◽  
Vol 15 (5) ◽  
pp. 472-485 ◽  
Author(s):  
Kuo-Chen Chou ◽  
Xiang Cheng ◽  
Xuan Xiao

<P>Background/Objective: Information of protein subcellular localization is crucially important for both basic research and drug development. With the explosive growth of protein sequences discovered in the post-genomic age, it is highly demanded to develop powerful bioinformatics tools for timely and effectively identifying their subcellular localization purely based on the sequence information alone. Recently, a predictor called “pLoc-mEuk” was developed for identifying the subcellular localization of eukaryotic proteins. Its performance is overwhelmingly better than that of the other predictors for the same purpose, particularly in dealing with multi-label systems where many proteins, called “multiplex proteins”, may simultaneously occur in two or more subcellular locations. Although it is indeed a very powerful predictor, more efforts are definitely needed to further improve it. This is because pLoc-mEuk was trained by an extremely skewed dataset where some subset was about 200 times the size of the other subsets. Accordingly, it cannot avoid the biased consequence caused by such an uneven training dataset. </P><P> Methods: To alleviate such bias, we have developed a new predictor called pLoc_bal-mEuk by quasi-balancing the training dataset. Cross-validation tests on exactly the same experimentconfirmed dataset have indicated that the proposed new predictor is remarkably superior to pLocmEuk, the existing state-of-the-art predictor in identifying the subcellular localization of eukaryotic proteins. It has not escaped our notice that the quasi-balancing treatment can also be used to deal with many other biological systems. </P><P> Results: To maximize the convenience for most experimental scientists, a user-friendly web-server for the new predictor has been established at http://www.jci-bioinfo.cn/pLoc_bal-mEuk/. </P><P> Conclusion: It is anticipated that the pLoc_bal-Euk predictor holds very high potential to become a useful high throughput tool in identifying the subcellular localization of eukaryotic proteins, particularly for finding multi-target drugs that is currently a very hot trend trend in drug development.</P>


Author(s):  
Dhilsath Fathima.M ◽  
S. Justin Samuel ◽  
R. Hari Haran

Aim: This proposed work is used to develop an improved and robust machine learning model for predicting Myocardial Infarction (MI) could have substantial clinical impact. Objectives: This paper explains how to build machine learning based computer-aided analysis system for an early and accurate prediction of Myocardial Infarction (MI) which utilizes framingham heart study dataset for validation and evaluation. This proposed computer-aided analysis model will support medical professionals to predict myocardial infarction proficiently. Methods: The proposed model utilize the mean imputation to remove the missing values from the data set, then applied principal component analysis to extract the optimal features from the data set to enhance the performance of the classifiers. After PCA, the reduced features are partitioned into training dataset and testing dataset where 70% of the training dataset are given as an input to the four well-liked classifiers as support vector machine, k-nearest neighbor, logistic regression and decision tree to train the classifiers and 30% of test dataset is used to evaluate an output of machine learning model using performance metrics as confusion matrix, classifier accuracy, precision, sensitivity, F1-score, AUC-ROC curve. Results: Output of the classifiers are evaluated using performance measures and we observed that logistic regression provides high accuracy than K-NN, SVM, decision tree classifiers and PCA performs sound as a good feature extraction method to enhance the performance of proposed model. From these analyses, we conclude that logistic regression having good mean accuracy level and standard deviation accuracy compared with the other three algorithms. AUC-ROC curve of the proposed classifiers is analyzed from the output figure.4, figure.5 that logistic regression exhibits good AUC-ROC score, i.e. around 70% compared to k-NN and decision tree algorithm. Conclusion: From the result analysis, we infer that this proposed machine learning model will act as an optimal decision making system to predict the acute myocardial infarction at an early stage than an existing machine learning based prediction models and it is capable to predict the presence of an acute myocardial Infarction with human using the heart disease risk factors, in order to decide when to start lifestyle modification and medical treatment to prevent the heart disease.


2020 ◽  
Author(s):  
Joseph Prinable ◽  
Peter Jones ◽  
David Boland ◽  
Alistair McEwan ◽  
Cindy Thamrin

BACKGROUND The ability to continuously monitor breathing metrics may have indications for general health as well as respiratory conditions such as asthma. However, few studies have focused on breathing due to a lack of available wearable technologies. OBJECTIVE Examine the performance of two machine learning algorithms in extracting breathing metrics from a finger-based pulse oximeter, which is amenable to long-term monitoring. METHODS Pulse oximetry data was collected from 11 healthy and 11 asthma subjects who breathed at a range of controlled respiratory rates. UNET and Long Short-Term memory (LSTM) algorithms were applied to the data, and results compared against breathing metrics derived from respiratory inductance plethysmography measured simultaneously as a reference. RESULTS The UNET vs LSTM model provided breathing metrics which were strongly correlated with those from the reference signal (all p<0.001, except for inspiratory:expiratory ratio). The following relative mean bias(95% confidence interval) were observed: inspiration time 1.89(-52.95, 56.74)% vs 1.30(-52.15, 54.74)%, expiration time -3.70(-55.21, 47.80)% vs -4.97(-56.84, 46.89)%, inspiratory:expiratory ratio -4.65(-87.18, 77.88)% vs -5.30(-87.07, 76.47)%, inter-breath intervals -2.39(-32.76, 27.97)% vs -3.16(-33.69, 27.36)%, and respiratory rate 2.99(-27.04 to 33.02)% vs 3.69(-27.17 to 34.56)%. CONCLUSIONS Both machine learning models show strongly correlation and good comparability with reference, with low bias though wide variability for deriving breathing metrics in asthma and health cohorts. Future efforts should focus on improvement of performance of these models, e.g. by increasing the size of the training dataset at the lower breathing rates. CLINICALTRIAL Sydney Local Health District Human Research Ethics Committee (#LNR\16\HAWKE99 ethics approval).


2020 ◽  
Author(s):  
Mikołaj Morzy ◽  
Bartłomiej Balcerzak ◽  
Adam Wierzbicki ◽  
Adam Wierzbicki

BACKGROUND With the rapidly accelerating spread of dissemination of false medical information on the Web, the task of establishing the credibility of online sources of medical information becomes a pressing necessity. The sheer number of websites offering questionable medical information presented as reliable and actionable suggestions with possibly harmful effects poses an additional requirement for potential solutions, as they have to scale to the size of the problem. Machine learning is one such solution which, when properly deployed, can be an effective tool in fighting medical disinformation on the Web. OBJECTIVE We present a comprehensive framework for designing and curating of machine learning training datasets for online medical information credibility assessment. We show how the annotation process should be constructed and what pitfalls should be avoided. Our main objective is to provide researchers from medical and computer science communities with guidelines on how to construct datasets for machine learning models for various areas of medical information wars. METHODS The key component of our approach is the active annotation process. We begin by outlining the annotation protocol for the curation of high-quality training dataset, which then can be augmented and rapidly extended by employing the human-in-the-loop paradigm to machine learning training. To circumvent the cold start problem of insufficient gold standard annotations, we propose a pre-processing pipeline consisting of representation learning, clustering, and re-ranking of sentences for the acceleration of the training process and the optimization of human resources involved in the annotation. RESULTS We collect over 10 000 annotations of sentences related to selected subjects (psychiatry, cholesterol, autism, antibiotics, vaccines, steroids, birth methods, food allergy testing) for less than $7 000 employing 9 highly qualified annotators (certified medical professionals) and we release this dataset to the general public. We develop an active annotation framework for more efficient annotation of non-credible medical statements. The results of the qualitative analysis support our claims of the efficacy of the presented method. CONCLUSIONS A set of very diverse incentives is driving the widespread dissemination of medical disinformation on the Web. An effective strategy of countering this spread is to use machine learning for automatically establishing the credibility of online medical information. This, however, requires a thoughtful design of the training pipeline. In this paper we present a comprehensive framework of active annotation. In addition, we publish a large curated dataset of medical statements labelled as credible, non-credible, or neutral.


GigaScience ◽  
2021 ◽  
Vol 10 (5) ◽  
Author(s):  
Teng Miao ◽  
Weiliang Wen ◽  
Yinglun Li ◽  
Sheng Wu ◽  
Chao Zhu ◽  
...  

Abstract Background The 3D point cloud is the most direct and effective data form for studying plant structure and morphology. In point cloud studies, the point cloud segmentation of individual plants to organs directly determines the accuracy of organ-level phenotype estimation and the reliability of the 3D plant reconstruction. However, highly accurate, automatic, and robust point cloud segmentation approaches for plants are unavailable. Thus, the high-throughput segmentation of many shoots is challenging. Although deep learning can feasibly solve this issue, software tools for 3D point cloud annotation to construct the training dataset are lacking. Results We propose a top-to-down point cloud segmentation algorithm using optimal transportation distance for maize shoots. We apply our point cloud annotation toolkit for maize shoots, Label3DMaize, to achieve semi-automatic point cloud segmentation and annotation of maize shoots at different growth stages, through a series of operations, including stem segmentation, coarse segmentation, fine segmentation, and sample-based segmentation. The toolkit takes ∼4–10 minutes to segment a maize shoot and consumes 10–20% of the total time if only coarse segmentation is required. Fine segmentation is more detailed than coarse segmentation, especially at the organ connection regions. The accuracy of coarse segmentation can reach 97.2% that of fine segmentation. Conclusion Label3DMaize integrates point cloud segmentation algorithms and manual interactive operations, realizing semi-automatic point cloud segmentation of maize shoots at different growth stages. The toolkit provides a practical data annotation tool for further online segmentation research based on deep learning and is expected to promote automatic point cloud processing of various plants.


2021 ◽  
Vol 13 (9) ◽  
pp. 1779
Author(s):  
Xiaoyan Yin ◽  
Zhiqun Hu ◽  
Jiafeng Zheng ◽  
Boyong Li ◽  
Yuanyuan Zuo

Radar beam blockage is an important error source that affects the quality of weather radar data. An echo-filling network (EFnet) is proposed based on a deep learning algorithm to correct the echo intensity under the occlusion area in the Nanjing S-band new-generation weather radar (CINRAD/SA). The training dataset is constructed by the labels, which are the echo intensity at the 0.5° elevation in the unblocked area, and by the input features, which are the intensity in the cube including multiple elevations and gates corresponding to the location of bottom labels. Two loss functions are applied to compile the network: one is the common mean square error (MSE), and the other is a self-defined loss function that increases the weight of strong echoes. Considering that the radar beam broadens with distance and height, the 0.5° elevation scan is divided into six range bands every 25 km to train different models. The models are evaluated by three indicators: explained variance (EVar), mean absolute error (MAE), and correlation coefficient (CC). Two cases are demonstrated to compare the effect of the echo-filling model by different loss functions. The results suggest that EFnet can effectively correct the echo reflectivity and improve the data quality in the occlusion area, and there are better results for strong echoes when the self-defined loss function is used.


Sign in / Sign up

Export Citation Format

Share Document