AI Challenges Research Articles

Beyond Fairness Metrics: Roadblocks and Challenges for Ethical AI in Practice

Beyond Fairness Metrics: Roadblocks and Challenges for Ethical AI in Practice

Local and Global Feature-Aware Dual-Branch Networks for Plant Disease Recognition.

Accurate identification of plant diseases is important for ensuring the safety of agricultural production. Convolutional neural networks (CNNs) and visual transformers (VTs) can extract effective representations of images and have been widely used for the intelligent recognition of plant disease images. However, CNNs have excellent local perception with poor global perception, and VTs have excellent global perception with poor local perception. This makes it difficult to further improve the performance of both CNNs and VTs on plant disease recognition tasks. In this paper, we propose a local and global feature-aware dual-branch network, named LGNet, for the identification of plant diseases. More specifically, we first design a dual-branch structure based on CNNs and VTs to extract the local and global features. Then, an adaptive feature fusion (AFF) module is designed to fuse the local and global features, thus driving the model to dynamically perceive the weights of different features. Finally, we design a hierarchical mixed-scale unit-guided feature fusion (HMUFF) module to mine the key information in the features at different levels and fuse the differentiated information among them, thereby enhancing the model's multiscale perception capability. Subsequently, extensive experiments were conducted on the AI Challenger 2018 dataset and the self-collected corn disease (SCD) dataset. The experimental results demonstrate that our proposed LGNet achieves state-of-the-art recognition performance on both the AI Challenger 2018 dataset and the SCD dataset, with accuracies of 88.74% and 99.08%, respectively.

Unlocking Public Health Data: Navigating New Legal Guardrails and Emerging AI Challenges.

Here, we analyze the public health implications of recent legal developments - including privacy legislation, intergovernmental data exchange, and artificial intelligence governance - with a view toward the future of public health informatics and the potential of diverse data to inform public health actions and drive population health outcomes.

The Generative AI challenges for competition authorities

The Generative AI challenges for competition authorities

APPLICATION OF AI WRITING ASSISTANT SOFTWARE IN EFL WRITING IN CHINA: A REVIEW

Technological innovations and the need for mastery of the English language in China have contributed to the introduction of Artificial Intelligence (henceforth AI) in teaching and learning the English language. The rationality behind this article is the need to understand and optimize the acquisition and the teaching of English writing skill among China EFL students and teachers. The purpose of this review is two fold. First, author wanted to explore the features, benefits and challenges of AI utilized in teaching and learning English language. Next, this review also sought to explore the strength and limitations of AI-based writing assistant software. To do this, as systematic review was conducted concerning the application and research of AI in language education have been reviewed to investigate the varying contributions that AI has had in the teaching and learning of English language writing. Using Scopus database, a total of 84 peer-reviewed papers were identified and 21 were analyzed. It further analyses sites such as Pigai.org, Iwrite, and other valuable software to assist in determining the required results. An in-depth analysis and evaluation of the use of AI in the teaching of English writing skill seems to suggest that the use of automated writing evaluation (henceforth AWE) or automated essay scoring (henceforth AES) has greatly facilitated the tasks of English writing skill acquisition, the identification of the areas of improvement that need to be focused on in the process of student’s writing, as well as issues and challenges related to the use of online resources at colleges and universities in China to enable the teaching and learning of the English language, particularly writing skill to be executed more effectively.

Future research recommendations for transforming higher education with generative AI

Higher education is crucial for producing ethical citizens and professionals globally. The introduction of generative AI (GenAI), such as ChatGPT, has posed opportunities and challenges to the traditional model of education. However, the current conversations primarily focus on policy development and assessment, with limited research on the future of higher education. GenAI's impact on learning outcomes, pedagogy, and assessment is crucial for reforming and advancing the workforce. This qualitative study aims to investigate student perspectives on GenAI's impact on higher education. The study uses an initial conceptual framework driven by a systematic literature review to investigate the opportunities and challenges of AI in education. This framework serves as an initial data collection and analysis framework. A sample of 51 students from three research-intensive universities was selected for this study. Thematic analysis identified three themes and 10 subthemes. The findings suggest that future higher education should be transformed to train students to be future-ready for employment in a society powered by GenAI. They suggest new learning outcomes—skills in learning and teaching with GenAI, AI literacy—and emphasize the significance of interdisciplinarity and maker learning, with assessment focusing on in-class and hands-on activities. They recommend six future research directions – competence for future workforce and its self-assessment measures, AI literacy or competency measures, new literacies and their relationships, interdisciplinary teaching, Innovative pedagogies and their evaluation, new assessment and its acceptance.

Balancing the Equation: Investigating AI Advantages, Challenges, and Ethical Considerations in the Context of GPT-3, Natural Language Processing, and Researcher Roles

Artificial Intelligence's (AI) revolutionary capacity is exemplified in GPT-3, an advanced model in Natural Language Processing. GPT-3's potential to generate human-like text and facilitate language tasks is unparalleled. This abstract highlights GPT-3's diverse applications, advantages, challenges, and ethical considerations. It underscores its role in enhancing research quality, promoting creative writing, and aiding language translation. However, it cautions against biases, inaccuracies, and limitations. Embracing GPT-3 requires balancing academic integrity, enriching education, and employing rigorous security measures. Ultimately, GPT-3 propels AI research by bridging language gaps and shaping innovative AI applications.

Tomato Leaf Disease Identification by Restructured Deep Residual Dense Network

As COVID-19 spread worldwide, many major grain-producing countries have adopted measures to restrict their grain exports; food security has aroused great concern from various parties. How to improve grain production has become one of the most important issues facing all countries. However, crop diseases are a difficult problem for many farmers so it is important to master the severity of crop diseases timely and accurately to help staff take further intervention measures to minimize plants being further infected. In this paper, a restructured residual dense network was proposed for tomato leaf disease identification; this hybrid deep learning model combines the advantages of deep residual networks and dense networks, which can reduce the number of training process parameters to improve calculation accuracy as well as enhance the flow of information and gradients. The original RDN model was first used in image super resolution, so we need to restructure the network architecture for classification tasks through adjusted input image features and hyper parameters. Experimental results show that this model can achieve a top-1 average identification accuracy of 95% on the Tomato test dataset in AI Challenger 2018 datasets, which verifies its satisfactory performance. The restructured residual dense network model can obtain significant improvements over most of the state-of-the-art models in crop leaf identification, as well as requiring less computation to achieve high performance

Publics’ views on ethical challenges of artificial intelligence: a scoping review

AbstractThis scoping review examines the research landscape about publics’ views on the ethical challenges of AI. To elucidate how the concerns voiced by the publics are translated within the research domain, this study scrutinizes 64 publications sourced from PubMed® and Web of Science™. The central inquiry revolves around discerning the motivations, stakeholders, and ethical quandaries that emerge in research on this topic. The analysis reveals that innovation and legitimation stand out as the primary impetuses for engaging the public in deliberations concerning the ethical dilemmas associated with AI technologies. Supplementary motives are rooted in educational endeavors, democratization initiatives, and inspirational pursuits, whereas politicization emerges as a comparatively infrequent incentive. The study participants predominantly comprise the general public and professional groups, followed by AI system developers, industry and business managers, students, scholars, consumers, and policymakers. The ethical dimensions most commonly explored in the literature encompass human agency and oversight, followed by issues centered on privacy and data governance. Conversely, topics related to diversity, nondiscrimination, fairness, societal and environmental well-being, technical robustness, safety, transparency, and accountability receive comparatively less attention. This paper delineates the concrete operationalization of calls for public involvement in AI governance within the research sphere. It underscores the intricate interplay between ethical concerns, public involvement, and societal structures, including political and economic agendas, which serve to bolster technical proficiency and affirm the legitimacy of AI development in accordance with the institutional norms that underlie responsible research practices.

Legal aspects of artificial intelligence in the employment process

The introduction of artificial intelligence in all domains of life is the most transformative process in recent history. It is also a highly dynamic process, and due to the pace of technological development, a very limited legal framework is available to address issues of human rights, ethics, transparency, privacy, safety and accountability. During the last few years, artificial intelligence started to reshape employment processes. Positive aspects of the introduction of AI in the employment process are efficiency and quality in job matching, digitalisation and acceleration of the process, ability to process large data and match job seekers to available vacancy announcements, the alleviation of administrative burdens of employees of employment agencies and giving them strategic and innovative roles. All these are indispensable in present times when demographic challenges in European countries are leading to increased labour migrations and require changes in the recruitment process. The paper explores the current challenges of AI, i.e. how to achieve human-centred values and fairness of AI use during the employment process, preventing algorithmic bias and discriminatory application of AI tools. In order to harness the maximum benefits of AI, we need to develop a regulatory framework that would be enforceable, inclusive and adaptive (OECD), particularly knowing that most AI solutions are privately owned and developed for commercial purposes.

AI in the Higher Military Institutions: Challenges and Perspectives for Military Engineering Training

Recognizing the peculiarities of the current system of military education and considering the necessity of rapid modernization of military engineering training, HMIs need to implement innovative technologies to enhance the educational process. The purpose was to present a detailed analysis of the implementation of AI technologies while training future military engineering officers, to outline the existing strategies, and to develop possible strategies to enhance the educational process through AI technologies. To achieve the research purpose, we conducted open and closed-ended surveys among 154 instructors through five questionnaires to address the research questions. The answers were studied using conventional content analysis and statistical data processing. The results revealed basic directions for using AI in military engineering training and possible AI applications for the formation of professional competencies among future military engineering officers. But, meanwhile, the findings indicate that the process of military engineering training is facing several challenges complicating the implementation of AI-driven transformations. To overcome the existing challenges of AI and elaborate the applicable recommendations for the implementation of AI in the HMIs, we outlined the strategies for the enhancement of military engineering training through AI technologies.

Current challenges of the state-of-the-art of AI techniques for diagnosing brain tumor

Background: Brain is the control center of the human body, in recent time, different variety of brain diseases are being discovered. The brain disease diagnosis tools are becoming challenging and still an open area of research, application of AI in brain disease diagnosis has made disease prediction and detection more precise and accurate. Automated technologies for non-invasive analysis of brain images have become necessary, because disease of brain is fatal and are the cause of large number of deaths in developed countries. Brain tumor surgery augmented with AI can result in safer and more effective treatment. The knowledge gap between clinical and data science experts still presents significant challenges. This paper will review literatures related to current challenges of AI technologies for brain tumor diagnosis and suggest new directions of AI technologies for diagnosing brain tumour. A systematic search of major academic databases (such as Science Direct, IEEE explore digital Library, and Google scholar) was conducted to identify relevant studies published between 2015 and 2023. The search term used in this study include “Brain tumor Diagnosis”, “AI challenges in Brain tumor Diagnosis”, ‘AI techniques” and “AI challenges in medicine and future”. Studies were included if they utilized AI techniques for brain tumor diagnosis. The identified studies were evaluated for the key challenges they encountered in their diagnostic approaches. The Present study identified several challenges related to the application of AI techniques in brain tumour diagnosis. These challenges include: Interpretability and explainability, variations in tumour location, shape, and size which make accurate segmentation and classification difficult. Overall, the challenges in explaining brain tumor detection stem from the unique requirements and complexities of the healthcare domain, necessitating specialized techniques and approaches. This study summarizes the new directions for AI as (I) Data Hungry: Large, standardized, annotated data sets and excellent ground truth data are necessary for the development of accurate AI. (II) Radiomics: makes it possible to extract a vast number of quantitative features from intricate clinical imaging arrays and convert them into high-dimensional data that can be further processed to determine their relationship to the histological features of the tumor, which represent underlying genetic mutations and malignancy as well as grade, progression, response to therapy, and even overall survival (OS). (III) Black box: AI, for instance, is capable of predicting the best course of care for a patient, but it is unable to explain its reasoning. A trend toward easing this restriction is interpretable deep learning, (IV) Demonstrating the generalizability of deep learning applications and conducting external validation are two major obstacles. (V) There are knowledge gaps in clinical oncology that need to be filled in order to successfully integrate AI and maximize its effects. (VI) Several national professional bodies have started programs to bridge these knowledge gaps and advance the adoption of AI in oncology in response to these difficulties.

Overcome Certification Challenges of AI Based Airborne Systems Using FAA Overarching Properties

AbstractTechnologies such as Artificial Intelligence encompassing Machine Learning, Deep Learning techniques have proved to be a game changer in multiple sectors. The use of these technologies in Aviation will prove to ensure higher rigor in predictive maintainability, air traffic management, safety assurance and airworthiness. Current traditional regulatory and industry standards have proved to be efficient and effective in providing highest level of safety. However, Certification of Avionics systems is becoming challenging because of increasing complexity of avionics systems. Current regulations and standards are very prescriptive and inadequate to support new technological trends. Civil Aviation Authorities (CAAs) are working on establishing the regulatory guidance around these technologies. This paper addresses the new strategy of employing Federal Aviation Administration (FAA) “Overarching Properties” ‐ a framework for demonstrating that a product has the qualities required to perform effectively and safely with the purpose of simplifying the certification process. This paper also covers the possible scenarios for establishing the safety net concept to enhance the determinism and airworthiness of AI based systems.

An empirical study on the response of the energy market to the shock from the artificial intelligence industry

This paper makes the first attempt to look at the response of the energy market to the shock from the AI industry and the role of AI in the energy market risk spillover network. The innovation of the paper lies in adopting a quantile-based and time-varying framework in the empirical analysis. This enables us to reveal the cross-quantile and lead-lag correlations between AI and the energy market. We find that an extreme negative shock from AI tends to correspond to an extreme downward movement in crude oil, gasoline, gas oil, and clean energy markets and the shock response dissipates after 66 days. Meanwhile, the cross-quantile correlations are subject to structural changes once the extreme events arrive. As for the total spillover effect of the energy market, it increases by 5.34 % after AI is considered as a source of uncertainty in the risk spillover network dominated by the crude oil market. Thus, the shock from AI challenges the stability of the energy market. A close relationship between the AI industry and the clean energy market is also identified. We contribute to studying the relationship between the AI industry and the energy market from the perspective of the risk connectedness between the two fields.

Dynamically engineered multi-modal feature learning for predictions of office building cooling loads

This paper reports a new knowledge-driven engineered feature learning approach in response to the Global AI Challenge for Building E&M Facilities held by the Electrical and Mechanical Service Department (EMSD) of the Hong Kong SAR. The results were awarded with a Grand Prize by the competition organizer. A dynamically engineered multi-modal feature learning (DEMMFL) method is proposed for predicting the cooling load of two office buildings. The DEMMFL model is estimated with the Lasso-ridge regression and compared with other well-known methods such as the Lasso. The novel approach applies control system knowledge to engineer useful features and explore load patterns for multi-mode modeling. Deep learning methods including LSTM, GRU, and AutoGluon are implemented for automated machine learning and tested in parallel to compare the performance of the proposed model with existing methods. The proposed model is demonstrated to predict long-term cooling load most accurately using engineered features from weather information only.

The Challenges of AI in Anesthesiology.

The Challenges of AI in Anesthesiology.

UNDERSTANDING BAYES RULE: BAYESIAN NETWORKS IN ARTIFICIAL INTELLIGENCE

Bayes' Rule and Bayesian Networks are foundational elements of AI, enabling probabilistic reasoning and informed decision-making in uncertain domains. This article introduces the core concepts and practical applications of these tools. We explore the historical origins, step-by-step construction of Bayesian Networks, and real-world AI applications. By understanding Bayes' Rule and Bayesian Networks, readers can unlock their potential to tackle complex AI challenges and uncertainties. And this article underscores the undeniable importance of Bayes' Rule and Bayesian Networks in AI. We hope to inspire a deeper understanding of these foundational concepts and encourage the exploration of their vast potential in artificial intelligence. Keywords: Bayesian Networks, AI, probabilistic reasoning, informed decision-making

Challenges of AI driven diagnosis of chest X-rays transmitted through smart phones: a case study in COVID-19

Healthcare delivery during the initial days of outbreak of COVID-19 pandemic was badly impacted due to large number of severely infected patients posing an unprecedented global challenge. Although the importance of Chest X-rays (CXRs) in meeting this challenge has now been widely recognized, speedy diagnosis of CXRs remains an outstanding challenge because of fewer Radiologists. The exponential increase in Smart Phone ownership globally, including LMICs, provides an opportunity for exploring AI-driven diagnostic tools when provided with large volumes of CXRs transmitted through Smart Phones. However, the challenges associated with such systems have not been studied to the best of our knowledge. In this paper, we show that the predictions of AI-driven models on CXR images transmitted through Smart Phones via applications, such as WhatsApp, suffer both in terms of Predictability and Explainability, two key aspects of any automated Medical Diagnosis system. We find that several existing Deep learning based models exhibit prediction instability–disagreement between the prediction outcome of the original image and the transmitted image. Concomitantly we find that the explainability of the models deteriorate substantially, prediction on the transmitted CXR is often driven by features present outside the lung region, clearly a manifestation of Spurious Correlations. Our study reveals that there is significant compression of high-resolution CXR images, sometimes as high as 95%, and this could be the reason behind these two problems. Apart from demonstrating these problems, our main contribution is to show that Multi-Task learning (MTL) can serve as an effective bulwark against the aforementioned problems. We show that MTL models exhibit substantially more robustness, 40% over existing baselines. Explainability of such models, when measured by a saliency score dependent on out-of-lung features, also show a 35% improvement. The study is conducted on WaCXR dataset, a curated dataset of 6562 image pairs corresponding to original uncompressed and WhatsApp compressed CXR images. Keeping in mind that there are no previous datasets to study such problems, we open-source this data along with all implementations.

Medical Artificial Intelligence Information Disclosure on Healthcare Professional Involvement in Innovation: A Transactional Theory of Stress and Coping Model

Complicated medical AI innovation activities call for healthcare professionals with more practical experience, who can propose their functional requirements for AI technology and products and provide useful feedback. Medical AI information disclosure is closely related to healthcare professionals. It is critical but may be a mixed blessing in promoting their participation in medical AI innovation. Based on the transactional theory of stress and coping, this study explored the influential mechanism of medical AI information disclosure on healthcare professional involvement in AI innovation. To examine the research model, the 356 valid responses of Chinese healthcare professionals were collected by two-stage online survey. Structural equation modeling was used to examine the effect of information disclosure on healthcare professional involvement and the mediation effect of challenge appraisal and hindrance appraisal. Ordinary least squares regression was used to examine the moderation effect of subjective norm. Results indicated that medical AI information disclosure through challenge appraisal was positively related to healthcare professional involvement, whereas it through hindrance appraisal negatively affected healthcare professional involvement. Moreover, subjective norm of AI use moderated negatively the impacts of medical AI information disclosure and challenge appraisal but moderated positively the impacts of medical AI information disclosure and hindrance appraisal. This study discussed medical AI information disclosure as a kind of stressor and expanded the application of the transactional theory of stress and coping in the smart healthcare context. It enriched the associations amongst tech-stressor, appraisal outcomes and individuals’ coping strategies and contributed to the development of medical AI innovation

Dynamic layer-span connecting spiking neural networks with backpropagation training

Spiking Neural Network (SNN) is one of the mainstream frameworks for brain-like computing and neuromorphic computing, which has the potential to overcome current AI challenges, for example, low-power learning dynamic processes. However, there is still a huge gap in performance between SNN and artificial neural networks (ANN) in traditional supervised learning. One solution for this problem is to propose a better spiking neuron model to improve its memory ability for temporal data. This paper proposed a leaky integrate-and-fire (LIF) neuron model with dynamic postsynaptic potential and a layer-span connecting method for SNN trained using backpropagation. The dynamic postsynaptic potential LIF model allows the neurons dynamically release neurotransmitters in an SNN model, which mimics the activity of biological neurons. The layer-span connecting method enhances the long-distance memory ability of SNN. We also first introduced a cosh-based surrogate gradient for the backpropagation training of SNNs. We compared the SNN with cosh-based surrogate gradient (CSNN), CSNN with dynamic postsynaptic potential (Dyn-CSNN), layer-span connecting CSNN (Las-CSNN), and SNN model with all the proposed methods (DlaCSNN-BP) in three image classification and one text classification datasets. The experimental results exhibited that proposed SNN methods could outperform most of the previously proposed SNNs and ANNs in the same network structure. Among them, the proposed DlaCSNN-BP got the best classification performance. This result indicates that our proposed method can effectively improve the effect of SNN in supervised learning and reduce the gap with deep learning. This work also provides more possibilities for putting SNN into practical application.