A survey by Malaysian Communication and Multimedia Commission (MCMC) has shown that Internet users in Malaysia has increased up to 88.7% in year 2020 compared to 76.9% in year 2016 which is quite high increase in perc...
详细信息
In this paper we give a narrative review of multi-modal video-language (VidL) models. We introduce the current landscape of VidL models and benchmarks, and draw inspiration from neuroscience and cognitive science to p...
In this paper we give a narrative review of multi-modal video-language (VidL) models. We introduce the current landscape of VidL models and benchmarks, and draw inspiration from neuroscience and cognitive science to propose avenues for future research in VidL models in particular and artificial intelligence (AI) in general. We argue that iterative feedback loops between AI, neuroscience, and cognitive science are essential to spur progress across these disciplines. We motivate why we focus specifically on VidL models and their benchmarks as a promising type of model to bring improvements in AI and categorise current VidL efforts across multiple ‘cognitive relevance axioms’. Finally, we provide suggestions on how to effectively incorporate this interdisciplinary viewpoint into research on VidL models in particular and AI in general. In doing so, we hope to create awareness of the potential of VidL models to narrow the gap between neuroscience, cognitive science, and AI.
An Information retrieval system is a software system that provides access to books, journals, and other documents. The IR system provides this access according to a specific user query to retrieve the intended book, d...
详细信息
ISBN:
(数字)9798350363203
ISBN:
(纸本)9798350363210
An Information retrieval system is a software system that provides access to books, journals, and other documents. The IR system provides this access according to a specific user query to retrieve the intended book, document… etc. The better IR system is the system that produces better results in a standard testing medium that contains collections of documents, queries, and relevant judgment lists. Relevant judgment lists are used to match each document to its relevant query, and the result of the IR system is compared to this judgment list to evaluate its performance. Usually, judgment lists are built by specialists in the domains of the document subject. In this paper, we used a Machine learning algorithm to create a judgment list with the least human involvement, saving much time and effort. The algorithm that is used is K-nearest neighbor, and the method that was used is the inverse of the distance. The requirements for this method are a few human-matched documents forming a small relevant judgment list, which saves effort and time, as we said before. The experiments were done on a test collection: Medline, which was composed of 1033 documents with 30 queries. This experiment serves as a preliminary study conducted on a small test collection, providing a foundation for further refinement and validation on larger collections in future work.
Contemporary research, particularly when addressing the most significant transdisciplinary research challenges, cannot effectively be done without a range of skills relating to data management, data analysis, and cybe...
Contemporary research, particularly when addressing the most significant transdisciplinary research challenges, cannot effectively be done without a range of skills relating to data management, data analysis, and cyberinfrastructure (CI). These data and CI skills are common to all disciplines that conduct data-centric research. Research Data Science acts as a vital component of the scientific process. In a grassroots attempt to address this gap, the CODATA-RDA Schools of Research Data Science (SoRDS) was founded in 2016 to provide instruction on foundational data science and open research concepts to early career researchers in low and middle-income countries. This partnership between international collaborators has since 2016 provided this training to over 1000 early career researchers in 24 events in 10 countries worldwide. The most recent event was held at Georgia Institute of Technology and focused on health equity and included researchers from minority-serving institutions in the southeast United States. This paper covers the background of the SoRDS project along with organization and curriculum details. It also covers the transition of the events from a non-domain-centric curriculum to spotlighting biological and social health equity data and what we learned to make future health-related events more engaging and valuable to the attendees. It also looks toward future events that will serve international students studying health informatics and other data-centric disciplines.
The configuration of Artificial Neural Networks (ANNs) in the context of predictive modelling can provide considerable difficulty owing to the complex nature of their arrangements and the need for meticulous hyperpara...
The configuration of Artificial Neural Networks (ANNs) in the context of predictive modelling can provide considerable difficulty owing to the complex nature of their arrangements and the need for meticulous hyperparameter adjustment. The present study addresses the issue by proposing a more straightforward methodology for configuring Artificial Neural Networks (ANNs) through the R programming language. The methodology presented in this study offers a systematic and comprehensive framework, ensuring accessibility and simplicity of implementation. This approach aims to enhance the usability of Artificial Neural Networks (ANNs) for practitioners who need advanced machine learning knowledge. In order to demonstrate the applicability of the proposed methodology, a series of experiments were conducted on a case study in sustainable energy research. This study makes a valuable contribution to academic discipline by establishing a connection between artificial neural network (ANN) theory and its practical application. This study aims to enhance the accessibility of artificial neural network (ANN) setup and provide significant insights to further progress predictive modelling, specifically focusing on sustainable energy research.
As fossil fuel reserves diminish, the urgency for clean, renewable energy has grown, with wind energy emerging as a leading solution. However, optimizing wind farm layouts remains particularly challenging due to the c...
详细信息
ISBN:
(数字)9798350363203
ISBN:
(纸本)9798350363210
As fossil fuel reserves diminish, the urgency for clean, renewable energy has grown, with wind energy emerging as a leading solution. However, optimizing wind farm layouts remains particularly challenging due to the complex wake effect, which significantly reduces energy conversion efficiency. Traditional mathematical approaches struggle to effectively solve the Wind Farm Layout Optimization (WFLO) problem. In response, this study proposes a Novel Adaptive LSHADE Algorithm (A-LSHADE) designed to address WFLO with complex wake effect. The algorithm adaptively guides the search process by selecting the optimal individuals at each iteration. Comprehensive experiments were performed under three different wind scenarios, analyzing the impact of varying turbine quantities on conversion efficiency. The results demonstrate that the proposed LSHADE variant consistently outperforms six state-of-the-art competitors in all tested conditions, delivering superior conversion efficiency.
Weakly supervised object detection (WSOD) is a task that uses only image-level category labels to train an object detector. The most common weakly supervised object detection framework uses ‘argmax’ as a baseline to...
详细信息
ISBN:
(数字)9798350363203
ISBN:
(纸本)9798350363210
Weakly supervised object detection (WSOD) is a task that uses only image-level category labels to train an object detector. The most common weakly supervised object detection framework uses ‘argmax’ as a baseline to filter spatially adjacent pseudo-labels, resulting in poor-quality pseudo-labels. In addition, the pseudo-label screening method based on IoU will lose many potential high-quality labels. To address the above problems, we propose a novel weakly supervised object detection method. We design a pseudo-label benchmark determination method, called area mapping, to determine the pseudo-label benchmark through the feature distribution of image and region proposals and improve the overall quality of pseudo-labels. We further propose a similarity-based pseudo-label discovery strategy utilizing the spatial similarity between pseudo-labels to discover high-quality pseudo-labels. Experiments were conducted on public datasets and new state-of-the-art results were obtained on VOC07 and surpassed the current state-of-the-art methods with a slight advantage on the COCO14 dataset.
With the continuous expansion of the tourism industry, understanding and enhancing tourism service quality has become increasingly critical. This study introduces an innovative method for evaluating tourism service qu...
详细信息
ISBN:
(数字)9798350363203
ISBN:
(纸本)9798350363210
With the continuous expansion of the tourism industry, understanding and enhancing tourism service quality has become increasingly critical. This study introduces an innovative method for evaluating tourism service quality using large language models (LLMs) and aspect-based sentiment analysis of travelogue data. By leveraging the advanced capabilities of LLMs, this method enables nuanced sentiment recognition at the aspect level, which significantly enhances the understanding of customer sentiments towards various tourism services. Analyzing over 1,000 travelogues sourced from an online platform, this research reveals the strengths of using LLMs in processing large volumes of unstructured data, resulting in a richer, data-driven understanding of customer experiences. The insights derived from this study provide valuable information for industry stakeholders, empowering them to tailor their services more effectively and ultimately enhance overall tourist satisfaction.
Financial Technology (Fin Tech) has become a pivotal subject in both academic and industry discourse over the past decade, with over 115,000 scholarly articles in the Web of Science (WoS) database and millions of news...
详细信息
ISBN:
(数字)9798350363203
ISBN:
(纸本)9798350363210
Financial Technology (Fin Tech) has become a pivotal subject in both academic and industry discourse over the past decade, with over 115,000 scholarly articles in the Web of Science (WoS) database and millions of news articles on platforms like Yahoo Finance. However, the majority of literature has examined only one aspect of these discussions, which hinders the progress and market entry of new Fin Tech innovations, thereby restricting the potential for technological advancement and industry transformation. This paper innovatively proposes a combined literature analysis from both academic and market perspectives. A bibliometric analysis of 3,255 academic articles on Fin Tech from the WoS database, covering the period from 2015 to 2024, was conducted. Additionally, Natural Language Processing (NLP) techniques, including keyword extraction and sentiment analysis, were utilized to explore the characteristics of 1,208 news articles on FinTech, sourced from Yahoo Finance from 2020 to 2024. A comprehensive model was developed to explore the characteristics of keywords and their involvement in both academic and market contexts. This approach has uncovered some key themes related to FinTech and provides a comprehensive overview of the connections between academia and the market.
It is widely believed that progress towards advanced artificial intelligence (AI) systems will only occur when brain-inspired information processing systems are available to guide autonomous behaviour. The development...
详细信息
It is widely believed that progress towards advanced artificial intelligence (AI) systems will only occur when brain-inspired information processing systems are available to guide autonomous behaviour. The development of such autonomous systems (AS) requires cognitive entities capable of acquiring information, learning, planning, and reasoning that eventually adapt to environmental uncertainties. Cognitive dynamic systems (CDS) provide an engineering tool to design AS. Although the literature addresses various aspects of cognitive dynamic systems, it needs a holistic methodology to analyze alternative techniques concerning the driving module that internally guides the system toward a goal. Since uncertain events far from expectation play a significant role in triggering information-seeking behaviours, contextual surprise can be considered an intrinsic motivator in CDS. This presentation highlights connections and similarities among several definitions of contextual surprise. For demonstration purposes, it reviews the design of a linear Gaussian CDS as a motivating example. It uses a contextual surprise minimization scheme to express information utility and guide CDS’ state estimation and control.
暂无评论