Research Chinese-Urdu Machine Translation Based on Deep Learning

Urdu is Pakistan 's national language. However, Chinese expertise is very negligible in Pakistan and the Asian nations. Yet fewer research has been undertaken in the area of computer translation on Chinese to Urdu. In order to solve the above problems, we designed of an electronic dictionary for Chinese-Urdu, and studied the sentence-level machine translation technology which is based on deep learning. The Design of an electronic dictionary Chinese-Urdu machine translation system we collected and constructed an electronic dictionary containing 24000 entries from Chinese to Urdu. For Sentence we used English as an intermediate language, and based on the existing parallel corpus of Chinese to English and English to Urdu, we constructed a bilingual parallel corpus containing 66000 sentences from Chinese to Urdu. The Corpus has trained by using two NMT Models (LSTM,Transformer Model) and the above two translation model were compared to the desired translation, with the help of bilingual valuation understudy (BLEU) score. On NMT, The LSTM Model is gain of 0.067 to 0.41 in BLEU score while on Transformer model, there is gain of 0.077 to 0.52 in BLEU which is better than from LSTM Model score. Furthermore, we compared the proposed model with Google and Microsoft translation.

The QR code intelligent positioning system of the LBS cloud platform in the Internet of things environment

Aiming at the problems of long positioning time and poor positioning accuracy in traditional positioning systems, a WeChat applet QR code area positioning system based on the LBS cloud platform is proposed and designed. The overall architecture of the system is divided into three parts: LBS cloud service, central data processing, and QR code positioning terminal for small programs. The hardware is designed from the server-side module, processor and positioning module to provide a basis for system construction. In the software design, the WeChat applet QR code area image is collected, the image edge features are enhanced and filtered, the positioning target is determined according to the processed image edge features, and the WeChat applet QR code area positioning system design is completed. The experimental results show that the positioning time of the system is equivalent to 50% of the traditional system, and the positioning accuracy is always maintained above 99.5%, which has significant advantages.

Automatic Text Summarization for Urdu Roman Language by Using Fuzzy Logic

In the new era of technology, there is the redundancy of information in the internet world, which gives a hard time for users to contain the willed outcome it, to crack this hardship we need an automated process that riddle and search the obtained facts. Text summarization is one of the normal methods to solve problems. The target of the single document epitome is to raise the possibilities of data. we have worked mostly on extractive stationed text summarization. Sentence scoring is the method usually used for extractive text summarization. In this paper, we built an Urdu Roman Language Dataset which has thirty thousand articles. We follow the Fuzzy good judgment technique to clear up the hassle of text summarization. The fuzzy logic approach model delivers Fuzzy rules which have uncertain property weight and produce an acceptable outline. Our approach is to use Cosine similarity with Fuzzy logic to suppress the extra data from the summary to boost the proposed work. We used the standard Testing Method for Fuzzy Logic Urdu Roman Text Summarization and then compared our Machine-generated summary with the help of ROUGE and BLEU Score Method. The result shows that the Fuzzy Logic approach is better than the preceding avenue by a meaningful edge.

Prediction of Building Energy Consumption Based on IPSO-CLSTM Neural Network

Accurate prediction of building load is essential for energy saving and environmental protection. Exploring the impact of building characteristics on heating and cooling load can improve energy efficiency from the design stage of the building. In this paper, a prediction model of building heating and cooling loads is proposed, which based on Improved Particle Swarm Optimization (IPSO) algorithm and Convolution Long Short-Term Memory (CLSTM) neural network model. Firstly, the characteristic variables are extracted and evaluated by Spearman’s correlation coefficient method; Then the prediction model based on the CLSTM neural network is constructed to predict building heating and cooling load. The IPSO algorithm is adopted to solve the problem that manual work cannot precisely adjust parameters. In this method, the optimization ability of the PSO algorithm is improved by changing the updating rule of inertia weight and learning factors. Finally, the parameters of the neural network are taken as IPSO optimization object to improve the prediction accuracy. In the experimental stage of this paper, a variety of algorithm models are compared, and the results show that IPSO-CLSTM can get the best results in the prediction of heating and cooling load.

Deep Learning and Autoregressive Approach for Prediction of Time Series Data

Fractional This paper proposes a deep learning approach for prediction of nonstationary data. A new regression scheme has been used in the proposed model. Any non-stationary data can be used to test the efficiency of the proposed model, however in this work stock data has been used due to the fact that stock data has a property of being nonlinear or non-stationary in nature. Beside using proposed model, predictions were also obtained using some statistical models and artificial neural networks. Traditional statistical models did not yield any expected results; artificial neural networks resulted into high time complexity. Therefore, deep learning approach seemed to be the best method as of today in dealing with such problems wherein time complexity and excellent predictions are of concern.

An Obstacle Avoidance Approach Based on Naive Bayes Classifier

Obstacle avoidance plays an important role in mobile robot. However, the traditional methods of obstacle avoidance have difficulty in distinguishing multiple obstacles by edge detection. In this paper, the traditional obstacle avoidance methods are improved to realize the function of multi-obstacle avoidance. Regarding the implementation process, the LiDAR is used instead of the RGBD camera, which reduces the difficulty of handling image noise and achieves reliable obstacle detection. It can accurately detect the borders of the nearest obstacle even in complex environments and perform obstacle avoidance. Regarding the obstacle avoidance prediction, the model training is performed through the Naive Bayes classifier based on the three attributes of the velocity of the robot, the left boundary of the obstacle and the right boundary of the obstacle. In the training process, dataset was expanded to enhance the accuracy of classifier model. When the robot goes forward, the improved method enable the robot to move at a higher velocity. The results show the feasibility of advanced obstacle avoidance method by simulation.

Skin Lesion Border Detection Based on Best Statistical Model Using Optimal Colour Channel

This paper proposes an effective way to segment melanoma skin lesion in colour dermoscopic images, using an edge-based approach. The proposed method, different methods were combined to improve the segmentation performance. These methods are morphological operations, bilateral filter, spline, polynomial model and canny edge detector. Different methods were tested to select the best method that was produced the best outcome. These testing methods, bilateral filter provided the highest PSNR amongst other filters such as median filter, Gaussian and average filter. Two statistical models were implemented polynomial model and linear regression and selected the best performance as polynomial model. Four edge detectors were applied to detect the edge of skin lesion and select the best segmentation accuracy. Manual border selection was used as the benchmark to evaluation the accuracy of the automatic border. The proposed method was able to achieve a good average accuracy of 96.69% based on canny edge detector. Our dataset consists of (70) dermoscopic images that includes melanoma and nevus.

INDECISIVENESS IN PROBLEMS OF MACHINE LEARNING AND AI

The problem of indecisiveness is integral part in each scientific research. However, it is still not a certainty whether this problem has an objective nature. In this paper we will extend the analysis of the sources and causes of indecisiveness and define the new categories that are a stumbling block in writing high quality software. Based on a sample, we will propose several ways to classify indecisiveness. Specifically, we will investigate indecisiveness related to a human, machine and environment. In some cases, it is possible to distinguish between remediable and unavoidable indecisiveness depending on the cause.

A Fractional PID Controller Based on Particle Swarm Optimization Algorithm

Fractional PID controller is a convenient fractional structure that has been used to solve many problems in automatic control. The fractional scale proportional-integral-differential controller is a generalization of the integer order PID controller in the complex domain. By introducing two adjustable parameters and , the controller parameter tuning range becomes larger, but the parameter design becomes more complex. This paper presents a new method for the design of fractional PID controllers. Specifically, the parameters of a fractional PID controller are optimized by a particle swarm optimization algorithm. Our simulation results on cold rolling APC system show that the designed controller can achieve control accuracy higher than that of a traditional PID controller.

A Study of Neural Machine Translation from Chinese to Urdu

Machine Translation (MT) is used for giving a translation from a source language to a target language. Machine translation simply translates text or speech from one language to another language, but this process is not sufficient to give the perfect translation of a text due to the requirement of identification of whole expressions and their direct counterparts. Neural Machine Translation (NMT) is one of the most standard machine translation methods, which has made great progress in the recent years especially in non-universal languages. However, local language translation software for other foreign languages is limited and needs improving. In this paper, the Chinese language is translated to the Urdu language with the help of Open Neural Machine Translation (OpenNMT) in Deep Learning. Firstly, a Chineseto Urdu language sentences datasets were established and supported with Seven million sentences. After that, these datasets were trained by using the Open Neural Machine Translation (OpenNMT) method. At the final stage, the translation was compared to the desired translation with the help of the Bleu Score Method.