Field Of Artificial Intelligence Research Articles

The Use of Deepfake Technology in China: Problems of Legal Regulation and Ways to Solve Them

The core of the deepfake technology is based on a generative-adversarial network built on a combination of two neural networks: a generative network (network G) creates samples, a discriminative network (network D) tries to distinguish correct («genuine») samples from incorrect ones. Networks G and D compete with each other thousands or even millions of times until network G improves its performance. Thus, the network D will no longer be able to distinguish real data from fake data. With the development of big data and machine learning technologies, the scenario for using deepfake technology has gradually changed from creating sound models and imitating text to deep video forgery. For a long time, images modified using traditional Photoshop and other technologies were easily recognized. Deepfake technology changed this situation, making it more difficult to identify fakes. As an important technological innovation in the field of artificial intelligence, deepfake technology is widely used in various areas of society, creating enormous applied value. However, any technology is a double-edged sword. The use of deepfake technology poses a great threat to personal privacy, property security and even national security. In order to find a balance between technological innovation and risk prevention and control, countries around the world are actively exploring various ways to manage. The paper describes the main risks posed by modern deepfake technology, provides an overview of legal regulation in this area in China and offers an effective way to solve problems.

Deriving comprehensive literature trends on multi-omics analysis studies in autism spectrum disorder using literature mining pipeline

Autism spectrum disorder (ASD) is characterized by highly heterogenous abnormalities in functional brain connectivity affecting social behavior. There is a significant progress in understanding the molecular and genetic basis of ASD in the last decade using multi-omics approach. Mining this large volume of biomedical literature for insights requires considerable amount of manual intervention for curation. Machine learning and artificial intelligence fields are advancing toward simplifying data mining from unstructured text data. Here, we demonstrate our literature mining pipeline to accelerate data to insights. Using topic modeling and generative AI techniques, we present a pipeline that can classify scientific literature into thematic clusters and can help in a wide array of applications such as knowledgebase creation, conversational virtual assistant, and summarization. Employing our pipeline, we explored the ASD literature, specifically around multi-omics studies to understand the molecular interplay underlying autism brain.

Impact of basic artificial intelligence (AI) course on understanding concepts, literacy, and empowerment in the field of AI among students

AbstractWith the development of information technologies and information processing methods, it is important to provide high‐quality education in the field of artificial intelligence (AI). The study aims to investigate the impact of an educational course on AI on the comprehension of concepts, literacy, and empowerment in the field of AI among students of higher educational institutions. The experiment involved 125 students from Hohai University in China. As a result of taking the training course, students were able to improve their understanding of concepts (increasing their average score from 6.33 to 9.69), literacy (from 2.94 to 3.99), and empowerment (from 3.90 to 4.04) in AI. The resulting data statistically confirmed the effectiveness of the developed course for improving confidence in the field of AI. The training module can be applied to improve confidence in the field of AI for students in various careers, as information competence is important these days and increases the success of graduates in employment. When it comes to further research, the encouraging results of this study suggest opportunities for promoting this training program among a diverse group of participants. To confirm the effectiveness of the developed course, it can be conducted among students in schools and other educational institutions, reducing it to even more basic if necessary.

2D-SnS-Embedded Schottky Device with Neurotransmitter-Like Functionality Produced Using Proximity Vapor Transfer Method for Photonic Neurocomputing.

Neuromorphic computing, which involves the creation of artificial synapses capable of mimicking biological brain activity, has intrigued researchers in the field of artificial intelligence (AI). To advance neuromorphic computing, a highly efficient 2D material-based artificial synapse capable of performing logical and arithmetic operations must be developed. However, fabricating large, uniform films or high-quality structures of 2D materials remains challenging because of their multistep and complex fabrication processes. In the present study, to produce large (Ø ≈ 3 in.), uniform, transparent neuromorphic devices, a novel single-step approach called proximity vapor transfer (PVT) that utilizes van der Waals (vdW) materials is employed. This single-step technique, which involves the fabrication of vdW materials on various substrates (glass, ITO, AZO, Mo, and Cu), allows control of the thickness and bandgap tunability. The Schottky device developed via the PVT method using vdW SnS with neurotransmitter (acetylcholine)-like functionality emulates biological synapses and exhibits photoelectronic synaptic behavior with wide-field-of-view synaptic plasticity. In addition, logic gate operations (NOT, OR, AND), reward-cascade neurotransmission, and imaging can be performed using 3 × 3 arrays of the device. This study represents a significant step toward the development of transparent and large-area synaptic devices, which are crucial for advancing AI applications.

Machine Learning and AI Innovations with Python: Trends and Future Directions

This comprehensive article examines the current state and future directions of machine learning and artificial intelligence innovations using Python. It provides an in-depth analysis of recent advancements in popular Python-based libraries such as TensorFlow, PyTorch, and scikit-learn, highlighting key features and performance improvements. The article explores emerging machine learning techniques, including federated learning, few-shot learning, and explainable AI, discussing their principles, Python implementations, and real-world applications. Furthermore, the article investigates future trends in AI development with Python, considering potential technological advancements, new libraries and frameworks, and emerging research areas such as quantum machine learning and neurosymbolic AI. It also addresses the challenges and opportunities facing Python in the evolving landscape of AI, including performance optimization, ethical considerations, and the increasing demand for production-ready AI solutions. By synthesizing current developments and future prospects, this article offers valuable insights for researchers, developers, and organizations seeking to leverage Python's capabilities in the rapidly advancing field of artificial intelligence and machine learning.

An Investigation Into To The Use Of Artificial Intelligence In The Indian Tax System

Indian taxation is main source of public finance in developing economy. The Indian tax system has long been plagued by issues including tax evasion and ineffective administration. Tax administration systems must constantly be designed to reduce errors and speed up decision-making. The Indian tax system is beset by a shortage of workers to do labor-intensive duties including data input, return inspection, tax audits, etc. The Indian government recently announced the use of artificial intelligence and machine learning in the tax assessment system in order to handle the evolving tax landscape in conjunction with the use of analytics. In the field of taxes, artificial intelligence, or AI, is a relatively recent development.

Data mining in AI: Evolution, applications, and future directions

Abstract. This paper provides a comprehensive analysis of the evolution and impact of data mining in the field of artificial intelligence (AI), with a particular focus on its application within social and information networks. It traces the origins of AI back to the 1956 Dartmouth Conference, highlighting the subsequent advancements in technologies such as machine learning and data mining that have fueled AI's growth. The paper explores the multifaceted applications of data mining in various sectors including healthcare, transportation, and industrial manufacturing, and delves into the challenges and innovations in recommendation systems, matrix factorization, and intelligent control of autonomous vehicles in intelligent transportation systems. The study emphasizes the significance of distributed algorithms and big data processing frameworks in enhancing the efficiency and applicability of data mining techniques.

THE ROLE OF EXPERT SYSTEMS WITHIN ARTIFICIAL INTELLIGENCE

Expert systems (ES) arose as a significant practical result in the application and development of artificial intelligence methods - a set of scientific disciplines that study methods for solving problems of an intellectual (creative) nature using a computer. From the beginning of its development, the field of artificial intelligence has considered several very complex problems, which, along with others, are still the subject of research: automatic theorem proofs, machine translation, image recognition, planning, game algorithms, strategies, etc. In modern understanding, an expert system is a set of programs that perform the functions of an expert when solving problems from a certain subject area. Expert systems advise, conduct analysis, provide consultations, and diagnose. The practical use of expert systems in enterprises contributes to work efficiency and improved qualifications of specialists. When creating expert systems, several difficulties arise. This is primarily because the customer cannot always accurately formulate his requirements for the system being developed. It is also possible that difficulties of a purely psychological nature may arise when creating a knowledge base of a system, an expert may hinder the transfer of his knowledge, fearing that he will subsequently be replaced by a “machine”. But these fears are not justified, since expert systems are incapable of learning, they do not have common sense or intuition. Currently, expert systems are being developed that implement the idea of self-learning. Advantages of an ES over a human expert. Knowledge-based systems have certain advantages over human experts: they have no prejudices, they don't rush to conclusions, these systems work systematically, looking at all the details, often choosing the best alternative from all possible ones,the knowledge base can be very, very large, once entered the machine, the knowledge is stored forever, a person has a limited knowledge base, and if data is not used for a long time, then it is forgotten and lost forever. Knowledge-based systems are resistant to “interference.” The expert uses collateral knowledge and is easily influenced by external factors that are not directly related to the problem being solved. ES that are not burdened with knowledge from other areas is, by their nature, less susceptible to “noise.” Over time, knowledge-based systems may be viewed by users as a type of replication - a new way of recording and disseminating knowledge. Like other types of computer programs, they cannot replace a person in solving problems but rather resemble tools that enable him to solve problems faster and more efficiently. These systems do not replace a specialist but are a tool in his hands. The modern idea of expert systems is given in the article. Their differences from traditional software products are shown, and advantages and disadvantages are considered. A conclusion is drawn about the perspectives of development. Keywords: information technology, expert systems, knowledge base, knowledge processing algorithm.

Bibliometric Analysis of Artificial Intelligence Studies in Computational Thinking

Computational thinking is an approach used to solve complex problems with basic skills such as decomposition, pattern recognition, abstraction and algorithm design. Especially artificial intelligence and its sub-fields make significant contributions to the development of these skills. This study aims to evaluate the research in the field of artificial intelligence in the computational thinking process through bibliometric analysis. The data of the study were obtained from the Web of Science (WoS) database on May 1, 2024. The VOSviewer software program, which is frequently used in bibliometric analysis and can be used free of charge, was used to analyze the data obtained from the WoS database. Analyses using the data obtained from the WoS database enabled the analysis of research in this field in terms of countries, authors, co-authorship status, publication journals, publication years and keywords. The findings of the study show that Taiwan is the country with the highest number of studies in this field, while Spain is the country with the highest number of citations. Hooshyar, D and Hsu, T are among the authors who contributed to the most publications. Collaboration plays an important role in research and Lai, C stands out as the author with the most co-authorship. Education and Information Technologies and Kunstliche Intelligenz are among the journals where the most studies are published. As a result of the examination of the 5 most cited studies, it was determined that different research methods and participant groups were used, data collection tools varied and different data analysis methods were used. These studies provide a broad perspective on artificial intelligence studies in the field of computational thinking and make important contributions to the literature in this field.

Feature attributions for water–solubility predictions obtained via artificial intelligence methods and chemists

Abstract The field of explainable artificial intelligence (AI) has garnered significant research interest. In particular, “feature attribution” in the field of chemistry has been focused upon. However, studies on comparisons of the relationship between AI- and human-based feature attributions when predicting the same outcome are scarce. Thus, this study aimed to investigate this relationship by comparing machine-learning-based feature attributions (graph neural networks and integrated gradients) with those of chemists (Hansch–Fujita method) considering the case of predicting water–solubility. The AI-based attributions were found to be similar to those of chemists despite their distinct origins.

The future of warfare. Anticipated changes in military trends

Purpose: to discuss the changes in the scientific and technical field, and examine their impact on military art. Method: comparative analysis, and synthesis. Findings: Today's military and political realities require the Armed Forces to be ready for wars that may occur at any time, including in the future, in order to ensure the country's national security. Because, as in the past, the struggle for natural resources, competition between states, aspirations for hegemony, as well as ideological differences and security concerns, which lead to military conflicts, will continue in the future. At the same time, the great progress achieved and expected in the field of technology will significantly change the traditional forms and methods of fighting. The study and skillful application of these forms and methods in the Armed Forces will be one of the main conditions for success in future wars. Scientific and technical progress affects all areas of human activity without exception, and has a noticeable effect on the development of methods and forms of war. In the future, the content and typology of military operations, strategy and tactics, forms and methods of application of armed forces will be formed under the influence of scientific and technical progress, mainly in the fields of artificial intelligence, biotechnology, and nanotechnology. For this reason, in order to study the nature of future wars, to predict the methods and forms of military operations, the article considers the causes of wars, discusses the changes in the scientific and technical field, and examines their impact on military art. Theoretical implications: The paper enhances our understanding of significant transformations due to advancements in technology, changes in geopolitical dynamics, and evolving societal expectations, against the backdrop of the extensive use of autonomous systems, including drones, unmanned vehicles, and AI-powered decision-making tools. Practical implications: The paper contibutes to the development of military strategies and operational plans, which consider how technological, social, and geopolitical changes will concretely affect the future warfare. Value: The study offers significant value in understanding and predicting the nature and character of future warfare and provides multiple lenses through which to analyze the war and devise strategies in line with emerging security threats.

Using Explainable Artificial Intelligence to Interpret Remaining Useful Life Estimation with Gated Recurrent Unit

In engineering, prognostics can be defined as the estimation of the remaining useful life of a system given current and past health conditions. This field has drawn attention from research, industry, and government as this kind of technology can help improve efficiency and lower the costs of maintenance in a variety of technical applications. An approach to prognostics that has gained increasing attention is the use of data-driven methods. These methods typically use pattern recognition and machine learning to estimate the residual life of equipment based on historical data. Despite their promising results, a major disadvantage is that it is difficult to interpret this kind of methodologies, that is, to understand why a certain prediction of remaining useful life was made at a certain point in time. Nevertheless, the interpretability of these models could facilitate the use of data-driven prognostics in different domains such as aeronautics, manufacturing, and energy, areas where certification is critical. To help address this issue, we use Local Interpretable Model-agnostic Explanations (LIME) from the field of eXplainable Artificial Intelligence (XAI) to analyze the prognostics of a Gated Recurrent Unit (GRU) on the C-MAPSS data. We select the GRU as this is a deep learning model that a) has an explicit temporal dimension and b) has shown promising results in the field of prognostics and c) is of simplified nature compared to other recurrent networks. Our results suggest that it is possible to infer the feature importance for the GRU both globally (for the entire model) and locally (for a given RUL prediction) with LIME.

Artificial intelligence in clinical practice: Quality and evidence

A revolution is taking place within the field of artificial intelligence (AI) with the emergence of generative AI. Although we are in an early phase at the clinical level, there is an exponential increase in the number of scientific articles that use AI (discriminative and generative) in their methodology. According to the current situation, we may be in an "AI bubble" stage; requiring filters and tools to evaluate its application, based on the quality and evidence provided. In this sense, initiatives have been developed to determine standards and guidelines for the use of discriminative AI (CONSORT AI, STARD AI and others), and more recently for generative AI (the CHART collaborative). As a new technology, AI requires scientific regulation to guarantee the efficacy and safety of its applications, while maintaining the quality of care; an evidence-based AI (IABE).

Comparative studies "Effect of educational policy on artificial intelligence in the countries of America, Singapore and Iran

Artificial intelligence under the paradigm of the fourth industrial revolution has become one of the fateful trends of societies in such a way that university education methods, as one of the most important social constructions, have influenced the advancement of technology. Undoubtedly, educational policy has changed the speed, depth and extent of artificial intelligence technology Artificial intelligence is one of the most important emerging technologies, which today has caused huge changes in all areas, including educational policymaking, at a high speed. Among the important applications of artificial intelligence in educational policymaking, we can mention the prioritization of issues based on the real needs and demands of the society and the diagnosis of problems based on the existing conditions, which will lead to the design and formulation of macro-educational policies. The use of methods such as neural networks, natural language processing, and machine learning algorithms makes the process of policy making more intelligent and dynamic, and as a result, the role of policymakers in the application of this technology becomes more efficient. Today, new technologies such as artificial intelligence, sensor networks, blockchain are changing aspects of everyday life. The study of artificial intelligence began in the 1940s and in recent years it has overshadowed the government departments of a large number of countries, of which e-government is one of the examples. One of the topics that is analyzed in the discussion of educational policy is the effect of content on artificial intelligence. In other words, understanding the complexities of decision-making processes as well as the importance of the actors involved in such processes, has undoubtedly influenced artificial intelligence and is influenced by it. This article is designed as a review and library method and it examines the educational policies of Iran, America and Singapore in the field of artificial intelligence with an exploratory view.

Qualitative and Quantitative Handling of Collective Labor Disputes and Strikes With the Help of Artificial Intelligence: an Inter- and Trans-disciplinary Theoretical Approach to Modern Industrial Relations

In today’s context of rapid technological evolution, this article explores from a purely theoretical perspective how artificial intelligence (AI) can be used as a tool for qualitative and quantitative manipulation of collective labor disputes and strikes. Beyond the simple use of modern technologies, this article examines the theoretical and epistemological foundations of integration (IA) in the analysis and prediction of labor conflicts. A theoretical framework is proposed that combines the theory of social conflict and the theory of industrial relations with contemporary approaches in the field of artificial intelligence, thus offering a new way to conceptualize and understand the dynamics of labor relations in the global economy. In conclusion, the paper discusses the epistemological, ethical and social implications of such an interdisciplinary approach.

Robust visual question answering via polarity enhancement and contrast

The Visual Question Answering (VQA) task is an important research direction in the field of artificial intelligence, which requires a model that can simultaneously understand visual images and natural language questions, and answer questions related to images. Recent studies have shown that many Visual Question Answering models rely on statistically regular correlations between questions and answers, which in turn weakens the correlation between visual content and textual information. In this work, we propose an unbiased Visual Question Answering method to solve language priors from the perspective of strengthening the contrast between the correct answer and the positive and negative predictions. We design a new model consisting of two modules with different roles. We input the image and the question corresponding to it into the Answer Visual Attention Modules to generate positive prediction output, and then use a Dual Channels Joint Module to generate negative prediction output with great linguistic prior knowledge. Finally, we input the positive and negative predictions together with the correct answer to our newly designed loss function for training. Our method achieves high performance (61.24%) on the VQA-CP v2 dataset. In addition, most existing debiasing methods improve performance on VQA-CP v2 dataset at the cost of reducing performance on VQA v2 dataset, while our method not only does not reduce the accuracy on VQA v2 dataset. Instead, it improves performance on both datasets mentioned above.

Deep neural networks for knowledge-enhanced molecular modeling

Designing neural networks for molecular modeling is a crucial task in the field of artificial intelligence. The goal is to utilize neural networks to understand and design molecules, which has significant implications for drug development and other real-world applications. Recently, with the advancement of deep learning, molecular modeling has made considerable progress. However, current methods are primarily data-driven, overlooking the role of domain knowledge, such as molecular shapes, in the modeling process. In this paper, we systematically investigate how incorporating molecular shape knowledge can enhance molecular modeling. Specifically, we design two deep neural networks, ShapePred and ShapeGen, to utilize molecular shapes in molecule prediction and generation. Experimental results demonstrate that integrating shape knowledge can significantly improve model performance. Notably, ShapePred exhibits strong performance across 11 molecule prediction datasets, while ShapeGen can more efficiently generate high-quality drug molecules based on given target proteins.

A StarGAN and transformer-based hybrid classification-regression model for multi-institution VMAT patient-specific quality assurance.

The field of artificial intelligence (AI)-based patient-specific quality assurance (PSQA) for volumetric modulated arc therapy (VMAT) faces challenges in terms of developing general models across institutions due to the prevalence of multi-institution data collection and multivariate heterogeneity. Building a general model that is capable of handling diverse multi-institution data is critical for enabling large-scale integration and analysis. This study aims to develop a star generative adversarial network (StarGAN) and transformer-based hybrid classification-regression PSQA framework to address unification of heterogeneous data from different institutions. A StarGAN and transformer-based hybrid classification-regression model was developed as a general PSQA framework to predict gamma passing rates (GPRs) and classify quality assurance (QA) results as "Pass" or "Fail" at multiple institutions. A total of 1815 VMAT plans were collected from eight institutions to develop the general PSQA framework and perform clinical commissioning and implementation. Among them, 20 independent clinical plans from each of eight institutions, for a total of 160 plans, were used for the clinical commissioning, and 205 new clinical plans from eight institutions were used for clinical implementation. For the 3%/3, 3%/2, and 2%/2mm gamma criteria, the sensitivity of the proposed PSQA framework with pretraining was 90.13%, 92.03%, and 95.84%, respectively, while the specificity was 76.01%, 76.12%, and 85.34%, respectively. The mean absolute errors (MAEs) of the proposed PSQA framework with pretraining were 1.36%, 2.37%, and 3.96%, respectively, while the root-mean-square errors (RMSEs) were 2.31%, 3.89%, and 5.17%, respectively. The results demonstrated visible improvement at multiple institutions. For clinical commissioning, the deviations between the predicted and measured results were all within 3% for 3%/3 and 3%/2mm at eight institutions. For clinical implementation, all failure plans were correctly identified by the proposed PSQA framework. The general PSQA framework enables diverse clinical data sources to be handled to achieve enhanced model performance and generalizability, and provides a solution to the unification of heterogeneous data from different institutions to construct robust QA models. This approach can be clinically deployed for VMAT QA.

Exploring Generative Artificial Intelligence

This work reviews the concept of Generative Artificial Intelligence (GAI), its main characteristics, and its classification inside the Artificial Intelligence (AI) field. In this way, the main GAI foundational algorithmic architectures for data generation are presented; each considers a previously gathered data set and generates novel and synthetic data. Subsequently, use cases for generative models reported in the literature are presented, ranging from applications in healthcare, education, and customer experiences. The deployment of GAI models still needs to overcome a series of challenges inspected in this document, considering the presence of bias, lack of transparency, and hallucinations presented by the models. Finally, the risks and misuse of GAI are reviewed since widespread use of this type of technology is expected in society.

Trends in the Use of Artificial Intelligence in Science Education: Bibliometric & Biblioshiny Analysis (1975-2024)

This research was conducted to look at research trends in the Scopus database regarding the study of Artificial Intelligence in Science Education. Using the PRISMA method, researchers selected and determined 146 documents that were ready to be analyzed using bibliometric and biblioshiny methods with the help of the R program and Vosviewer. From the results of the analysis carried out, it was concluded that research on Artificial Intelligence in Science Education in the Scopus Database started from 1975 to 2024, with an 8-fold increase in the number of publications, and 13 times the number of citations in the last 7 years. China was the country with the highest contribution to this research, showing 17(11.64%) publications and 418(13.65%) citations. "BMC Medical Education" is the most productive source with an h-index of 4. Jose Carpio Cañada from Universidad de Huelva in Spain is the most productive author with an h-index of 2. Y. K. Dwivedi et al., (2021) with 1159(37.85%) citations is the source with highest citation. There are 58 keywords divided into 9 clusters, with the keywords "Digital Technologies", "ChatGPT", and "Systems Thinking" being the main recommendations in the field of artificial intelligence in science education