Proceedings of the 3rd International Conference on Signal Processing and Machine Learning
Omer Burak Istanbullu, Eskisehir Osmangazi University
The classic 2D SLAM are not good enough in nowadays environment. This report uses virtual machine with Ubuntu based Slam_bot package, based on the RTAB-MAP algorithm and Vision SLAM mapping to simulate the four-wheeled robot to autonomously navigate to the target point in various environments. Also, this report introduces a RGBD-SLAM based algorithm which combines the visual and depth data to process the data collect from the sensors. This robot has many sensors like, lidar sensors, RGB vision camera and odometry sensors. To see how the RTAB-MAP algorithm with RGB-D sensor replace for the 2D SALM. As results, the robot with RGB-D and RTAB-MAP algorithms have very good performance. The results show that the navigation system can complete the navigation and localization in lots of complex situations. However, some problems still exist, the speed of the robot is not fast. This may limit the application of the self-navigation robot to a certain extent, like some emergency occasion.
Based on the cooperation of artificial intelligence (AI) and unmanned driving technology, finding the best path from the starting node to the target node in the shortest time is a research hotspot. The required path planning algorithms can therefore be classified according to their different approaches to solving the problem. This paper focuses on Dijkstra’s algorithm, A* algorithm, Ant colony optimization, and genetic algorithm. It also discusses the present problems of these algorithms and some improvements made by researchers focusing on automated guided vehicle path planning.
Since the invention of the memristive device with a nano-scale footprint, a lot of scholars have started to focus on the area of recognition systems based on the Complementary Metal-Oxide Semiconductor chips (CMOS) integrated with memristive devices. This paper’s goal is to compare and analyze the advantage and disadvantage on the near research on the cognitive machine. Start with the construction of a simple dynamic model of neurons in Section 2, the history of the development of the recognitive machine is introduced in Section 3. Section 4 focusing on the comparison and analysis of researches done by different scholars. Through the comparison of multiple scholars’ work, improvement of the memory system would be a potential way to improve the recognition system nowadays, and it is discussed in the conclusion.
As the popularity of online social networks has increased, so has the manner in which people purchase online. Online reviews on the purchasing of items or the provision of services have become the primary source of user opinion. When a substantial body of knowledge assesses a product or service, that body's effect on the market is significant. Because of this, concerns have been raised among manufacturers and merchants, who frequently compose rubbish reviews to promote or denigrate the quality of specific items or services in order to make a profit or maintain their reputation. You may choose to promote or denigrate certain products or services that have been targeted. This kind of commentary is known as "trash remark." Comment spam violates the interests of enterprises to a significant level, and it possesses a defensive mechanism that combines the "Deepfake" of artificial intelligence technology, which renders conventional security measures inapplicable. Deep learning is the approach that will be most effective in resolving this issue. Researchers assess this research according on how they extract features from the commentary dataset and how they extract features from the commentary dataset, employing a variety of methodologies and strategies to find solutions to the problems. The authors also looked at the key deep learning technologies that have been suggested as a solution to the issue of spam detection in the primary machine, as well as the performance of various deep learning technologies. At the same time, the fact that there is not a lot of spam online review data makes this sort of research a highly significant one. The purpose of this article is to offer an insightful and all-encompassing analysis of comparative research on the topic of spam detection that has recently been conducted.
Semi-supervised learning is one of the potential research fields in text classification. In this paper, semi-supervised pseudo-label training experiments are conducted using the BERT model that has been pre-trained as a baseline. Only 20% of the original dataset is used for the new training set after segmenting the training set. The raw corpus used for pseudo-label training consists of the remaining 80% of data after labels are removed, while the original test set is still utilized. The results indicate that the key to the semi-supervised pseudo-labelling method is the performance of the original model and reasonable data filtering techniques. Even though the SoftMax value used for data filtering is not precisely equivalent to model prediction accuracy, experimental results show it can somewhat reduce the error propagation problem of the model. This is consistent with earlier research. However, using SoftMax as the threshold for data screening can't bring enough benefits to the model training and make it surpass the training performance of the original data set. As a result, future studies will focus on improving the accuracy of pseudo-labelling with a more suitable data selection method to better the model's performance.
With the widening gap between the rich and the poor, social inequality has emerged in all aspects. The issue of inequity in education needs more attention, because it is related to the development of national quality. There have been many studies on the use of online housing advertisements for feature extraction and semantic analysis, and the use of machine learning methods to construct models to predict socioeconomic status. This study considers the influence factors of education, and conducts Bayesian classification and LDA model analysis on all reviews of New York schools on the largest school rating and recording website in the United States to explore the primary factors associated with educational imbalance in a society. Results show that various requirements for teachers, such as teaching ability and student management ability, were the most important factors that appeared in the reviews. Gender issues are also very important in education. In terms of the overall parts of speech, the emotions are all positive, indicating that the current level of education can satisfy parents as a whole. However, there are still many potential problems of educational inequality that need to be discovered and solved, and the methods of inquiry need to be expanded and upgraded.
Urban crime poses a serious challenge to urban sustainability and livability. Many studies have been conducted to explore the patterns and causes of urban crime, as well as prevention techniques. Studies have found that neighborhood socioeconomic status affects the incidence of urban crime, but studies on this topic are limited due to data limitations. To fill this gap, this study designed an approach for Brooklyn, USA, that collects publicly available data from housing advertising sites and the Open Street Map and trains a machine learning model to predict fine-grained neighborhood socioeconomic status. The experimental results show that the gradient boosting decision tree regression model has the best prediction accuracy. Then, we verified the predicted significant correlation between fine-grained neighborhood socioeconomic status and criminal activity in the precinct by using a geographically weighted regression model, that is, criminal activity has a higher incidence in disadvantaged neighborhoods. It was also found that neighbourhood socioeconomic status was the best predictor of harassment and burglary.
The universal issue for authorities is to plan land use more effectively and efficiently, and to provide more sustainable mobility in urban areas. The sustainable prism model is proposed to achieve the requirements for sustainable development. Transit-oriented Development (TOD) is a city planning method that coordinates the mass transit system and the land use pattern. This article analyzes the sustainable prism and discusses the effect of TOD on sustainabilities in ecology, economy and equality to provide the city planner insight into the urban development pattern. Generally, TOD can reduce the energy consumption in the transport sector and related infrastructure, contribute to the air quality from the environmental aspect, facilitate employment and attract investment in the economic aspect and improve discrepancy in employment availability and transport burden among different groups from the social equality aspect. However, the noise pollution in the environment, the value capture effect in the economic aspect and the TOD efficiency in the social justice aspect are insufficiently discussed.
Urban vitality is defined as the zing of cities providing citizens with the ability to live, and it is important to offer an essential basis for estimating urban growth and spatial balance. Shenzhen’s rapid development has made remarkable achievements in a brief period of forty years as a special economic zone since reform and openness. This study takes Shenzhen as an example, uses 21 POI data obtained from Baidu Maps, and discusses the relationship between urban vitality and time and space and the influence factors of urban vitality. Results show that urban vitality has a close relationship with time and space and high-vitality areas are judged based on the production activities of urban people. Besides, amusement activities do not have a clear influence on urban vitality. Furthermore, convenient transportation and ordered life-supporting services play a stressful role in pushing urban vitality. Based on the above three elements, it is necessary to pay attention to taking production activities, creating convenient transportation, and providing ordered life-supporting services to increase a city’s urban vitality.
Diabetes is one of the most diseases in the world. In the last 40 years, the number of persons worldwide with diabetes has tripled. There were 108 million patients over the age of 18 in 1980 and 422 million in 2014, accounting for 8.5% of the entire population at that time. Diabetes directly caused 1.5 million fatalities worldwide in 2012, with hyperglycemia-related illnesses accounting for 2.2 million deaths. Diabetes is expected to be the 7th greatest cause of death by 2030 according to the World Health Organization. As the risk of diabetes increases, machine learning algorithms are used to improve early diagnosis of diabetes, and various researchers have also done some corresponding algorithms for predicting diabetes machine learning. As a commonly used machine learning algorithm, AdaBoost integrated learning algorithm is superior in the diagnosis and prediction of diabetes mellitus. In this paper, it is proposed that a hybrid model to detect the risk of diabetes. This hybrid model is detected and eliminated by K-means-based outliers, synthesizing the distribution of minority data oversampling techniques (SMOTE), and Adaboost to classify diabetes. According to the final experimental result, the model prediction accuracy is 0.950 after using the hybrid model in the PIMA dataset. In the future, if a larger number of sample training data are utilized for training, the model's accuracy will improve.
Style transfer is a wide-used technique in image and photograph processing, which could transfer the style of an image to a target image that has a different content. This image processing technique has been used in the algorithms of some image processing software as well as modern artistic creation. However, the intrinsic principle of style transfer and its transfer accuracy is still not clear and stable. This article discusses a new method for preprocessing image data that uses feature extraction and forming vector fields and utilizing multiple VGG19 to separately train the distinct features in images to obtain a better effect in predicting. Our model could generate more autonomous and original images that are not simply adding a style filter to the image, which can help the development of AI style transfer and painting.
In recent years, users’ requirements for the richness of car functions have gradually increased. The application of a large number of electronic components in the automobile makes the degree of complexity of the automobile higher, so the user has a higher demand for the stability of the automobile system. Especially once a key functional component fails, driving safety will be seriously affected. Because in the design process, we find and analyze the possible causes of failure, so as to avoid the impact of failure in the design as much as possible. We use FMEA (Failure Mode and Effect Analysis) Analysis method and FTA (Fault Tree Analysis) analysis method to analyze the car headlight fault. These two analysis methods can efficiently and accurately analyze faults and determine the most basic events leading to failure. Based on the above analysis results, the proposed solution to improve the system can be given, so that the system can meet the functional safety requirements.
In the case of acute stroke patients, mere diagnosis falls short, and segmentation is needed. Recent development in deep learning and image processing has provided us with the potential to automatically perform brain lesion segmentation. However, many of the approaches ended up failing to generalize to new data by overfitting the ATLAS R1.2 dataset and ignoring information extraction of the high-level features. We propose a novel Residual H-Net that addresses these two issues by adding a special residual block in the middle of the U-Net and increasing dilation size to better extract the high-level features. The presented model is less susceptible to overfitting and much easier to train. The Residual H-Net is tested on a subset of ATLAS R2.0 data and shows promising performance against the previous state-of-the-art model.
With the continuous development of facial expression technology, especially the development of deep learning, and the establishment of in-the-wild datesets in recent years, field face recognition has become a hot research field in the wild of facial expression recognition. Different from the traditional facial expression recognition(FER), in-the-wild facial facial expression recognition, the problem of recognition accuracy caused by illumination, occlusion or low image resolution. Therefore, in order to solve these problems, new methods have been put forward continuously in recent years. In this paper, we first continue to summarize the widely used datesets, and then summarize the paper methods of facial expression recognition in the field proposed in the past two years.
At the end of 2019, the pandemic began to have a huge impact on people's lives around the world. In order to control and prevent the virus from spreading, world-wide governments have encouraged all residents to wear masks in their daily lives. Since the 1960s, face recognition technology has gradually emerged and developed. In the 21st century, it has been widely used in security, finance, transportation, and even retail, advertising, smart devices, education, health care, entertainment and other fields, playing a pivotal role in everyone's life. In the era of COVID-19, the public wearing face masks has undoubtedly hindered the smooth technology application of face recognition. The technology of recognizing human face in the case of wearing masks should be developed, followed by the adjustment of the original technology and data. The development of technology also brings certain social impact. Our research reviews the adjustment and progress of face recognition technology in the era of COVID-19, summarizes the impact of technological progress on society and people's lives, and provides help for the government to better prevent and control the epidemic.
When making decisions, probabilistic reasoning is used to utilize the known information to predict or determine those unobserved factors and variables that are crucial for the outcome. To find the relationship between outcome and one or more predictors, linear regression is commonly used in statistics and machine learning algorithms. In this article, the basic concept of linear regression and hypothesis testing are reviewed. The common modern methods for variable and model selection including Stepwise selection, Akaike’s information criterion, Bayesian information criterion, and Mallow’s C_p are discussed and reviewed. Each method and criterion have its own uniqueness and limitation depending on the dataset and the purpose of analysis. Through discussing this, this paper aims to inform and explain each different method for variable and model selection in linear regression and provide information to help choose the most suitable methods to predict or find the relationship between response variables and the independent variables for analysis.
In our daily life, artificial intelligence technology has penetrated all aspects, for us, language communication is a kind of media, but also an important way for us to convey ideas and give orders, so now, language communication has become an important way for us to interact with artificial intelligence. This paper through social research and the use of voice command to the existing speech recognition technology products test method puts forward the current development problems and solutions, the current situation of artificial intelligence speech recognition technology and the future in the field of communication systems and the intelligent home application made a simple introduction.
The advancement of science and technology, as well as the passage of time, have resulted in an increase in the use of high-tech products in our daily lives, with microelectronics playing an important role as a fundamental technology. Today, the microelectronics industry is the world's sunrise industry. China is rapidly expanding its microelectronics industry. Microelectronics technology is one of the world's fastest growing technologies, and it is the foundation of the information industry in the information age. Now, microelectronics technology has become a standard for measuring a country's level of science and technology. This paper, using a method of literature review, focuses on discussing and summarizing the development and application of microelectronics technology in China, as well as providing some possible correct solutions.
The machine room inspection robot, as a special robot, can locate and navigate in the machine room independently, and carry sensors to detect equipment faults and record instrument data instead of the machine room staff. In this paper, the relevant standards and existing inspection robots at home and abroad were investigated and the overall scheme was determined based on the actual machine room environment. The monitoring background employs B/S architecture, allowing staff to deliver inspection tasks and manage robots from any computer. The motion map is created using the GMapping algorithm, and the topology map for inspection is created based on the inspection task and feasible path. For raster map localization, the adaptive Monte Carlo localization algorithm is used. The functional layer software for the robot is designed based on the inspection function to be performed by the robot and the man-machine interaction with the monitoring background. The findings of this paper's research indicate that the inspection robot system can meet the requirements of computer room autonomous inspection.
Virtual reality technology is an emerging technology has been used in many fields. Because of its unique multi-perception, interactivity and autonomy, can enable users to obtain an immersive experience, enrich the sense of experience, VR is promoting the development and creation of multiple industrial fields with high speed. This article partly summarizes the application of VR technology in medical, educational and cultural fields. This paper analyzes the application characteristics and status quo of VR in the medical field, and discusses the progress and advantages of VR in education. It also summarizes the related technology of VR in culture. With the successful research and development of 5G technology, VR technology will also be improved, so VR technology will also be more widely used in various fields and promote the development of various fields.