Artificial Intelligence Learning Algorithms Research Articles

Exploring functional conservation in silico: a new machine learning approach to RNA-editing.

Around 50years ago, molecular biology opened the path to understand changes in forms, adaptations, complexity, or the basis of human diseases through myriads of reports on gene birth, gene duplication, gene expression regulation, and splicing regulation, among other relevant mechanisms behind gene function. Here, with the advent of big data and artificial intelligence (AI), we focus on an elusive and intriguing mechanism of gene function regulation, RNA editing, in which a single nucleotide from an RNA molecule is changed, with a remarkable impact in the increase of the complexity of the transcriptome and proteome. We present a new generation approach to assess the functional conservation of the RNA-editing targeting mechanism using two AI learning algorithms, random forest (RF) and bidirectional long short-term memory (biLSTM) neural networks with an attention layer. These algorithms, combined with RNA-editing data coming from databases and variant calling from same-individual RNA and DNA-seq experiments from different species, allowed us to predict RNA-editing events using both primary sequence and secondary structure. Then, we devised a method for assessing conservation or divergence in the molecular mechanisms of editing completely in silico: the cross-testing analysis. This novel method not only helps to understand the conservation of the editing mechanism through evolution but could set the basis for achieving a better understanding of the adenosine-targeting mechanism in other fields.

THE TEACHING OF DOUBLE EFFECT AND ARTIFICIAL INTELLIGENCE

This paper explores the possibility of applying existing ethical principles to artificial intelligence learning algorithms. The processes controlled by artificial intelligence will have a significant impact on the lives of all people in the future. Decision-making without consideration of moral and ethical principles may lead to catastrophic consequences not only as direct damage, but also in the form of degradation of certain ethical norms in general. An attempt to directly transfer universal moral principles to the ethics of artificial intelligence may lead to insurmountable obstacles due to the nature of artificial intelligence.

Construction of an urban multi variable geospatial information integration model based on distributed artificial intelligence

In order to improve the ability of urban multi-spatial information integration, a multi-spatial information integration model based on artificial intelligence is proposed. The empirical analysis model of urban multivariate geospatial information integration is established, and the statistical characteristics are analysed according to the information distribution. The multi-resolution clustering model of urban multi-spatial information is established by using the multi-regression parameter method. The methods of fuzzy feature extraction and associated feature mining were used to extract information features, and the constraint parameters of urban multivariate geospatial information integration were analysed by GIS parameter estimation and statistical analysis. The parameter optimisation model of urban multivariate geospatial information integration is established and the artificial intelligence learning algorithm is used to integrate urban multivariate geospatial information. The simulation results show that this method has a good adaptability to the integration of urban multi-spatial information, and is helpful to the spatial distribution planning of geographic information, and improves the ability of accurate mining and scheduling of urban multi-spatial information.

Transition towards a full self-sufficiency through PV systems integration for sub-Saharan Africa: a technical approach for a smart blockchain-based mini-grid

Access to affordable, reliable and clean energy is an important sustainability goal of the United Nations. In areas where the public electricity grid is unreliable or unavailable, photovoltaic systems can be a solution. However, they are cost-intensive, mainly because of the energy storage systems. Mini-grids can be an answer for reducing upfront investment and overall system lifetime costs while increasing electricity availability. The mini-grid technology is mature, nevertheless, there are downsides when it comes to integrating existing solar systems of different manufacturers. The system topology is usually predefined and a central instance controls the mini-grid. Thus, the integration of existing power systems is difficult due to the communication constraints of these systems with the mini-grid controller. Including existing power systems into a decentralized mini-grid, can highly increase cost-efficiency. In a decentralized approach payments for the consumed energy between mini-grid actors are required. Accounting is, however, a complex administrative procedure, if the respective power systems are owned by different individuals and organizations. A transparent blockchain-based temper-proof approach can be a solution to automate metering and billing, allowing automatic payments between independent subsystem owners using smart contracts. In order to further optimize the smart mini-grid, an artificial intelligence learning algorithm for a dynamic electricity price needs to be developed. This smart and decentralized approach for building Mini-Grids is a novelty bringing solar systems one step closer to self-sufficiency. This paper describes how a smart mini-grid solution can be implemented using the Don Bosco Solar & Renewable Energy Center campus mini-grid in Tema, Ghana as a case study.

Guest Editorial: Special Issue on Artificial Intelligence for Robotics

The papers in this special section focus on artificial intelligence (AI) applications for robotics. AI technologies, covering cognition, analysis, inference, and decision-making, enable robots to act smartly and greatly enhance robots’ capabilities to assist and support humans. As a primary carrier of AI technologies, robotics is one of the applications for AI to demonstrate its strong ability. Augmenting robots with AI technologies for the engineering systems, robots are expected to have more pervasive applications in industry, agriculture, logistics, medicine, and to name but a few. This special issue will be dedicated to AI for robotics, including AI to support human–robot interaction, AI for multirobot systems, AI learning algorithms in robotics, ethics of AI in the context of robotics, and applications of AI-enabled robots. These papers offer innovative ideas and concepts, discoveries and improvements, and novel applications to the field of AI for robotics.

Space robot motion path planning based on fuzzy control algorithm

Path planning technology is one of the core technologies of intelligent space robot. Combining image recognition technology and artificial intelligence learning algorithm for robot path planning in unknown space environment has become one of the hot research issues. The purpose of this paper is to propose a spatial robot path planning method based on improved fuzzy control, aiming at the shortcomings of path planning in the current industrial space robot motion control process, and based on fuzzy control algorithm. This paper proposes a fuzzy obstacle avoidance method with speed feedback based on the original advantages of the fuzzy algorithm, which improves the obstacle avoidance performance of space robot under continuous obstacles. At the same time, we integrated the improved fuzzy obstacle avoidance strategy into the behavior-based control technology, making the avoidance become an independent behavioral unit. Divide the path planning into a series of relatively independent behaviors such as fuzzy obstacle avoidance, cruise, trend target, and deadlock by the behavior-based method. According to the interaction information between the space robot and the environment, each behavior acquires the dominance of the robot through the competition mechanism, making the robot complete the specific behavior at a certain moment, and finally realize the path planning. Furthermore, to improve the overall fault tolerance of the space, robot we introduced an elegant downgrade strategy, so that the robot can successfully complete the established task in the case of control command deterioration or failure of important information, avoiding the overall performance deterioration effectively. Therefore, through the simulation experiment of the virtual environment platform, MobotSim concluded that the improved algorithm has high efficiency, high security, and the planned path is more in line with the actual situation, and the method proposed in this paper can make the space robot successfully reach the target position and optimize the spatial path, it also has good robustness and effectiveness.

Intelligent Decision-Making System for Martial Arts Competition Using Deep Learning

In the field of martial arts, athletes can win the initiative in the competition if they can correctly and timely acquire the field knowledge, evaluate the situation efficiently, and formulate a suitable strategy. In this paper, we use fuzzy mathematics, mathematical statistics, and artificial intelligence learning algorithms to carry out systematic and in-depth research on the selection of Wushu competition scene decision-making. The fuzzy mathematics theory is combined with the intelligent design theory for decision-making based on a multiagent, case-based reasoning selection, and adaptability evaluation analysis. The Wushu competition scene decision system is constructed based on artificial intelligence learning algorithms. Our approach outperforms the existing approaches in terms of accuracy, sensitivity, specificity, and Matthew’s correlation coefficient (MCC). The results of our proposed model can be anticipated to have the potential for better flexibility and scalability in martial arts competition.

Bursting potentiates the neuro-AI connection.

For decades, researchers have wondered whether algorithms used by artificial neural networks might be implemented by biological networks. Payeur et al. have strengthened the connection between neuroscience and artificial intelligence by showing that biologically plausible mechanisms can approximate key features of an essential artificial intelligence learning algorithm.

Research on intelligent Diagnosis method of Electric vehicle Charging Fault based on artificial intelligence expert system

To realize charging electric vehicle safety, this paper studies the latest artificial intelligence learning algorithm, combined with the characteristics in the process of electric vehicle charging, based on the deep learning algorithm of charge-discharge failure correlation method, combining with the characteristics of the expert system, design of electric vehicle charging process of the fault intelligent diagnosis system structure, characteristic information database, knowledge base, reasoning machine, and for battery, charging pile, power supply equipment failure correlation between the integration of online diagnosis, which can effectively improve the electric car battery safety, promote the healthy development of electric vehicles.

Point-by-point feature extraction of artificial intelligence images based on the Internet of Things

In the context of the rapid development of artificial intelligence in the Internet of Things, the establishment of the Internet of Things can promote rapid progress in the field of artificial intelligence. Traditional image detection methods use wavelet energy algorithms to divide background and edge noise, with poor resolution and low image detection accuracy. A series of problems such as slow detection speed and lack of image depth analysis exists. Aiming at the disadvantages of traditional methods, this study proposes the design of an artificial intelligence image detection system based on the Internet of Things, and uses intelligent artificial pixel feature collection technology to extract point-by-point feature of the image. This paper introducing artificial intelligence learning algorithms to the wheel detection in the workshop under the Internet of Things system, which can not only solve the problem of poor feature anti-interference and poor robustness in the traditional method, but also has important significance for the secondary development of the wheel detection system. The neural network can be used to classify wheel images while integrating other detection requirements, such as wheel defect detection and wheel number identification. The rich data resources and processing capabilities of the Internet of Things to perform feature analysis and feedback on the collected image pixels are utilized. The artificial intelligence of image synthesis module performs image conversion processing on the signal processes the feedback signal. The analysis results can complete image detection and complete artificial intelligence images. Through simulation experiments, it is proved that the design of artificial intelligence image detection system based on the Internet of Things has the advantages of high image detection rate, high recognition accuracy, stable operation, and efficient processing. The design idea has good application value.

An embedded system based on DSP platform and PCA-SVM algorithms for rapid beef meat freshness prediction and identification

Statistical and artificial intelligence learning algorithms such as Principal Component Analysis (PCA) and Support Vector Machine (SVM) are widely used in many meat quality assessment applications to classify and predict the freshness of beef meat. This paper presents the implementation of the PCA and SVM algorithms on an embedded system based on a Digital Signal Processor (DSP). A dataset of eighty-one hue, saturation, and intensity (HSI) beef meat images was used. The PCA is used as a projection and prediction model where the SVM is used for the classification and identification of the beef meat. Results obtained from PCA projection model on a desktop system using Matlab software show the projection of three groups which represent the degree of beef meat freshness during the days of cold storage. A perfect prediction of the new unknown samples was obtained by PCA prediction model. A successful rate 100% of classification and identification was obtained by SVM. The PCA and SVM methods were implemented on the C6678 multi-core DSP. The implementation results of these algorithms were similar as those obtained by Matlab software. The processing time of the algorithms measured on the embedded system was lower compared to the desktop system. The embedded platforms based on DSP as portable tools can be used to predict or identify the sample beef meat freshness anywhere and in real-time.

Research on Intrusion Detection System Based on Clustering Fuzzy Support Vector Machine

Introducing the artificial intelligence learning algorithm to solve the problem of network security is a focus of current research. We introduce the clustering algorithm into artificial intelligence learning algorithm and apply Fuzzy Support Machines to the intrusion detection. We put forward a method which is based on Fuzzy Support Machines. Then, we chose an appropriate RBF kernel function according to the characteristic of intrusion detection. And we get the intrusion detection algorithm based on Fuzzy Support Machines. The algorithm in this paper reduces the training time and improves the efficiency of the algorithm. Experimental results show that this method improves the fuzzy support vector machine training efficiency, and it is also very effective in intrusion detection. The first part of this paper is the introduction of the related problem. The second part is the concept of Fuzzy Support Vector Machine. The third part is the choice of the clustering center. The fourth part is the process of intrusion detection algorithm. The final part is the experiment.

Surface-enhanced laser desorption/ionization time-of-flight mass spectrometry: serum protein profiling in seminoma patients

Surface-enhanced laser desorption ionization time-of-flight mass spectrometry (SELDI-TOF MS) allows rapid protein profiling of complex biological mixtures. We analyzed testicular germ cell cancer serum samples to differentiate between cancer and controls with a special focus on beta-hCG-negative seminomas. Proteomic spectra were generated by the ProteinChip system and analyzed by the proteomic platform "proteomic.net". For statistical analysis, an artificial intelligence learning algorithm was used. The classification algorithm correctly identified the pattern in 90.4% of the patients. Decision trees predicted seminomas with 91.5% sensitivity and 89.4% specificity. Seminoma patients with normal beta-hCG serum level were correctly predicted with 80% sensitivity and 70% specificity. Our study demonstrates protein profiles of testicular germ cell cancer patients that differ in a highly significant degree from normal controls. Validation of these findings may enable proteomic profiling to become a valuable tool, especially for aftercare.

SERUM PROTEOMIC PROFILING BY SURFACE-ENHANCED LASER DESORPTION/IONIZATION TIME-OF-FLIGHT MASS SPECTROMETRY IN TESTICULAR GERM CELL CANCER PATIENTS

Objectives: Surface-enhanced laser desorption ionization mass spectrometry (sELDI-TOF MS) incorporates time-of-flight protein mass spectrometry and chip-based chromatographic protein selection and allows rapid protein profiling of complex biological samples. In this study, serum protein profiles were examined in testicular germ cell cancer patients and healthy controls with the aim of predicting cancer patients by a molecular test combining on-chip chromatography and mass spectrometry. Methods: Serum samples were taken from 49 patients with histologically proven seminomas. Forty-nine healthy subjects served as controls. Mass-spectrometry-based SELDI-TOF ProteinChip technology was used for serum proteomic analysis. The samples were fractionated and bound to the IMAC-Cu+ and CM10 ProteinChip. Proteomic spectra were generated by the ProteinChip system and analyzed by the proteomic platform ?proteomic.net?. Serum samples were also examined for beta-human chorionic gonadotropin levels. An artificial intelligence learning algorithm was used to differentiate between cancer and control subjects. Results The classification algorithm correctly identified the pattern in 90.4% of the samples (95% confidence interval of 82.6%-95.5%). Decision trees predicted seminomas with 91.5% sensitivity and 89.4% specificity. Seminoma patients with normal beta-human chorionic gonadotropin serum levels (n=38 of 49) were correctly predicted with 80% sensitivity and 70% specificity. Conclusion Our study is the first to demonstrate that SELDI-TOF MS protein profiles of cancer and control subjects differ to a highly significant degree in testicular germ cell cancer patients. Validation of these findings may enable proteomic profiling to become a valuable tool, especially for aftercare.