检索结果-内蒙古大学图书馆

Sports-QA: A Large-Scale Video Question Answering Benchmark for Complex and Professional Sports

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Li, Haopeng Deng, Andong Liu, Jun Rahmani, Hossein Guo, Yulan Schiele, Bernt Bennamoun, Mohammed Ke, Qiuhong School of Computing and Information Systems University of Melbourne Australia Center for Research in Computer Vision University of Central Florida United States Pillar Singapore University of Technology and Design Singapore School of Computing and Communications Lancaster University United Kingdom School of Electronics and Communication Engineering Sun Yat-sen University China Department of Computer Vision and Machine Learning Max Planck Institute for Informatics Saarland Informatics Campus Germany School of Physics Maths and Computing University of Western Australia Australia Department of Data Science & AI Monash University Australia

Reasoning over sports videos for question answering is an important task with numerous applications, such as player training and information retrieval. However, this task has not been explored due to the lack of relevant datasets and the challenging nature it presents. Most datasets for video question answering (VideoQA) focus mainly on general and coarse-grained understanding of daily-life videos, which is not applicable to sports scenarios requiring professional action understanding and fine-grained motion analysis. In this paper, we introduce the first dataset, named Sports-QA, specifically designed for the sports VideoQA task. The Sports-QA dataset includes various types of questions, such as descriptions, chronologies, causalities, and counterfactual conditions, covering multiple sports. Furthermore, to address the characteristics of the sports VideoQA task, we propose a new Auto-Focus Transformer (AFT) capable of automatically focusing on particular scales of temporal information for question answering. We conduct extensive experiments on Sports-QA, including baseline studies and the evaluation of different methods. The results demonstrate that our AFT achieves state-of-the-art performance. Copyright © 2024, The Authors. All rights reserved.

关键词： Question answering

Virtual Reality in Manufacturing Education: A Scoping Review Indicating State-of-the-Art, Benefits, and Challenges Across Domains, Levels, and Entities

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Ipsita, Ananya Kaki, Ramesh Liu, Ziyi Patel, Mayank Duan, Runlin Deshpande, Lakshmi Yuan, Lin-Ping Lowell, Victoria Maharaj, Ashok Peppler, Kylie Feiner, Steven Ramani, Karthik School of Mechanical Engineering Purdue University West LafayetteIN47907 United States Department of Mechanical Engineering Birla Institute of Technology and Science Pilani Hyderabad Campus Telangana Hyderabad500078 India School of Electrical and Computer Engineering Purdue University West LafayetteIN47907 United States Learning Design and Technology Purdue University West LafayetteIN47907 United States Department of Computer Science Columbia University New YorkNY10027 United States School of Information and Computer Sciences University of California IrvineCA92697 United States School of Education University of California IrvineCA92697 United States Hong Kong University of Science and Technology Hong Kong TCS Avapresence- Spatial Intelligence Lab

To address the shortage of a skilled workforce in the U.S. manufacturing industry, immersive Virtual Reality (VR)-based training solutions hold promising potential. To effectively utilize VR to meet workforce demands, it is important to understand the role of VR in manufacturing education. Therefore, we conduct a scoping review in the field. As a first step, we used a 5W1H (What, Where, Who, When, Why, How) formula as a problem-solving approach to define a comprehensive taxonomy that can consider the role of VR from all relevant possibilities. Our taxonomy categorizes VR applications across three key aspects: (1) Domains, (2) Levels, and (3) Entities. Using a systematic literature search and analysis, we reviewed 108 research articles to find the current state, benefits, challenges, and future opportunities of VR in the field. It was found that VR has been explored in a variety of areas and provides numerous benefits to learners. Despite these benefits, its adoption in manufacturing education is limited. This review discusses the identified barriers and provides actionable insights to address them. These insights can enable the widespread usage of immersive technology to nurture and develop a workforce equipped with the skills required to excel in the evolving landscape of manufacturing. © 2025, CC BY.

关键词： Virtual environments

Bayesian Optimization Framework for Efficient Fleet design in Autonomous Multi-Robot Exploration

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Concha, David Molina Li, Jiping Yin, Haoran Park, Kyeonghyeon Lee, Hyun-Rok Lee, Taesik Sirohi, Dhruv Lee, Chi-Guhn The Dynamic Optimization and Reinforcement Learning Laboratory The Department of Mechanical & Industrial Engineering University of Toronto TorontoONM5S3G8 Canada The Department of Electrical & Computer Engineering University of Toronto TorontoONM5S3G8 Canada The Division of Engineering Science The Faculty of Applied Science and Engineering University of Toronto TorontoONM5S3G8 Canada The Complex System Design Laboratory The Department of Industrial and Systems Engineering Korea Advanced Institute of Science and Technology Daejeon34141 Korea Republic of The Department of Industrial Engineering Inha University Incheon22212 Korea Republic of

This study addresses the challenge of fleet design optimization in the context of heterogeneous multi-robot fleets, aiming to obtain feasible designs that balance performance and costs. In the domain of autonomous multi-robot exploration, reinforcement learning agents play a central role, offering adaptability to complex terrains and facilitating collaboration among robots. However, modifying the fleet composition results in changes in the learned behavior, and training multi-robot systems using multi-agent reinforcement learning is expensive. Therefore, an exhaustive evaluation of each potential fleet design is infeasible. To tackle these hurdles, we introduce Bayesian Optimization for Fleet design (BOFD), a framework leveraging multiobjective Bayesian Optimization to explore fleets on the Pareto front of performance and cost while accounting for uncertainty in the design space. Moreover, we establish a sub-linear bound for cumulative regret, supporting BOFD’s robustness and efficacy. Extensive benchmark experiments in synthetic and simulated environments demonstrate the superiority of our framework over state-of-the-art methods, achieving efficient fleet designs with minimal fleet evaluations. Copyright © 2024, The Authors. All rights reserved.

关键词： Multipurpose robots

Apparent vs. true battery aging: Impact of various load characteristics on accelerated aging tests

学校读者我要写书评

暂无评论

Journal of Energy Storage 2025年 127卷

作者： Schreiber, Markus Lin, Yilei Sommer, Alessandro Wassiliadis, Nikolaos Torricos, Pablo Morales Rogge, Marcel Lewerenz, Meinert Grosu, Cristina Endisch, Christian Jossen, Andreas Lienkamp, Markus School of Engineering & Design Department of Mobility Systems Engineering Institute of Automotive Technology Garching85748 Germany School of Engineering & Design Department of Mechanical Engineering Institute for Machine Tools and Industrial Management Garching85748 Germany Research Group Electromobility and Learning Systems Ingolstadt85019 Germany School of Engineering & Design Department of Energy and Process Engineering Institute of Electrical Energy Storage Munich80333 Germany

Vehicle applications involve highly dynamic and variable operating conditions. These result from changing driving profiles and diverse driving behaviors. To accelerate battery aging tests, experiments are often based on constant load cycles. While efficient, this approach distorts the natural degradation behavior of the cells. This study compares cell aging under real-world and laboratory load conditions. For this purpose, we compare realistic driving profiles measured in a test vehicle in the field with synthetic constant power cycles. Average discharge power and charging sequences are kept identical for comparability. Results show that constant power cycles cause stronger cell aging than dynamic loads. Differential voltage analysis indicates more lithium inventory loss and active material loss at the negative electrode under static loads. Static loading also leads to strong inhomogeneity in the negative electrode and altered charge transfer and diffusion kinetics, as shown by impedance data. After 15 months of rest, the cells recovered up to 52 % of lost capacity and 66 % of the resistance increase. Impedance assessment and differential voltage analysis confirm the recovery and rehomogenization of the negative electrode. These findings show that conventional accelerated tests are strongly influenced by relaxation and load dynamics. As a result, their applicability to specific use cases, such as automotive applications, is strongly limited. Therefore, andraditional test strategies require reconsideration and redesign. © 2025 The Authors

关键词： Negative resistance

Fast mapping in word-learning: A case study on the humanoid social robots' impacts on Children's performance

学校读者我要写书评

暂无评论

International Journal of Child-Computer Interaction 2023年 38卷

作者： Esfandbod, Alireza Rokhi, Zeynab Meghdari, Ali F. Taheri, Alireza Soleymani, Zahra Alemi, Minoo Karimi, Mahdie Social and Cognitive Robotics Laboratory Center of Excellence in Design Robotics and Automation (CEDRA) Sharif University of Technology Tehran Iran Fereshtegaan International Branch Islamic Azad University Tehran Iran Department of Speech Therapy School of Rehabilitation Tehran University of Medical Sciences Tehran Iran Department of Humanities West Tehran Branch Islamic Azad University Tehran Iran Mahan Learning Disorders Center Tehran Iran

The evolution of robotic technology has heightened the demands for robots capable of social interaction with humans, also known as social robots. The encouraging prospects of deploying this advanced assistive tool in entertainment, healthcare, and education have, in turn, promoted substantial growth in studies conducted on the human-robot interaction (HRI) topic. One hot HRI subject is the examination of social robots' efficacy within various educational scenarios, e.g., word learning. In the context of word learning, fast mapping (FM) is a well-established technique through which novel words are taught by contrasting them with previously known words. The concept of this approach originates from how infants comprehend new vocabularies' referents. In this study, we investigated the potential benefits of utilizing social robots during educational interventions, based on the FM strategy, by means of two measures, alternative forced-choice (AFC) test and gaze data analysis. The first metric explores participants' retention after a 10-min and a one-week delay, while the second is concerned with scrutinizing their gaze focus during training sessions. Our findings indicate that the utility of social robots during FM interventions could enhance children's engagement, attention, and retention of novel words. © 2023 Elsevier B.V.

关键词： Child-robot interaction (CRI) Fast mapping Social robots Word-learning

Medical Image Classification Using Thermal Images and Diagnosis of Carcinogen

学校读者我要写书评

暂无评论

Medical Image Classification Using Thermal Images and Diagno...

Innovation and Novelty in Engineering and technology (INNOVA), International Conference on

作者： Govinda Rajulu. G L. Sharmila D. Venkatesan K Srinivas Nallam Krishnaiah K. Laxman Kumar Department of Computer Science and Engineering Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology. Avadi Chennai Tamil Nadu India Dept. of Artificial Intelligence and Machine Learning St. Martin's Engineering College Hyderabad India Deptarment of CSE(AI & ML) St. Martin's Engineering College Hyderabad India Department of IT St. Martin's Engineering College Dhulapally Secunderabad Telangana India Deptartment of Computer Science and Design St. Martin's Engineering College Hyderabad India

ISBN: (数字)9798331505134

ISBN: (纸本)9798331505141

Lung and Colon (L&C) tumors are lethal sicknesses that can foster in a few organs all the while and, in specific circumstances, jeopardize human existence. Even though it is highly unlikely that these two types of cancer will develop simultaneously, delaying diagnosis significantly increases the likelihood of metastasis between the organs that are affected. To really treat specific sorts of malignant growth, histological determination is fundamental. In the past, doctors had to go through a long and difficult process to look at thermal images and figure out if a patient had cancer; however, this procedure may now be completed much more quickly thanks to the new technology options. Histological images of L&C tumors were classified using a hybrid Deep learning (DL) model with an attention mechanism and a multipath network in this study. To zero in on the main attributes and negligence the less significant ones, a consideration component was utilized. In a multipath network, data travels over a number of channels before each channel is converted and the output from all of the branches is combined. The multipath network is similar to grouped convolution when simplified. The five thermal image categories of the LC25000 dataset were utilized. Among these classes were two for colon disease and three for cellular breakdown in the lungs. The proposed model was compared to a number of well-known DL models, such as ResNet-50, VGG-16, and AlexNet. The proposed approach showed the best exhibition concerning exactness (99.2%), particularity (99.12%), responsiveness (99.28%), accuracy (99.12%), and F1 score (99.2%), as indicated by the exploratory discoveries

关键词： Technological innovation Accuracy Lungs Subject matter experts Medical services Colon Medical diagnostic imaging Cancer Tumors Testing

EXPANSE: A Deep Continual / Progressive learning System for Deep Transfer learning

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Iman, Mohammadreza Miller, John A. Rasheed, Khaled Branch, Robert M. Arabnia, Hamid R. Department of Computer Science Franklin College of Arts and Sciences University of Georgia AthensGA United States Institute for Artificial Intelligence Franklin College of Arts and Sciences University of Georgia AthensGA United States Learning Design and Technology Mary Frances Early College of Education University of Georgia AthensGA United States

Deep transfer learning (DTL) techniques try to tackle the limitations of deep learning, the dependency on extensive training data and the training costs, by reusing obtained knowledge from source data for target data. However, the current DTL techniques suffer from either catastrophic forgetting dilemma (losing the previously obtained knowledge) or overly biased pre-trained models (harder to adapt to target data) in finetuning pre-trained models or freezing a part of the pre-trained model, respectively. Progressive learning, a sub-category of DTL, reduces the effect of the overly biased model in the case of freezing earlier layers by adding a new layer to the end of a frozen pre-trained model. Even though it has been successful in many cases, it cannot yet handle distant source and target data. We propose a new continual/progressive learning approach for deep transfer learning to tackle these limitations. To avoid both catastrophic forgetting dilemma and overly biased-model problem, we expand the pre-trained model by expanding pre-trained layers (adding new nodes to each layer) in the model instead of only adding new layers. Hence the method is named EXPANSE. Our experimental results confirm that we can tackle distant source and target data using this technique. At the same time, the final model is still valid on the source data, achieving a promising deep continual learning approach. Moreover, we offer a new way of training deep learning models inspired by the human education system. We termed this two-step training: learning basics first, then adding complexities and uncertainties. The evaluation implies that the two-step training extracts more meaningful features and a finer basin on the error surface since it can achieve better accuracy in comparison to regular training. EXPANSE (model expansion and two-step training) is a systematic continual learning approach applicable to different problems and DL models. Copyright © 2022, The Authors. All rights reserved.

关键词： Freezing

AI-based sensor layout for predicting thermal deformations of CFRP machine tools

学校读者我要写书评

暂无评论

CIRP Annals 2025年

作者： Finkeldey, Felix Kato, Makoto Wiederkehr, Petra Kakinuma, Yasuhiro Virtual Machining Department of Computer Science TU Dortmund University Dortmund Germany Lamarr Institute for Machine Learning and Artificial Intelligence Dortmund Germany Department of System Design Engineering Faculty of Science & Technology Keio University Yokohama Kohoku-ku Japan

Using data-based approaches, accurate predictions of thermal deformations, which can significantly affect the quality of manufactured components, can be enabled. However, a sufficient amount of data with maximised information content is necessary for efficient training. In this paper, an approach for optimising sensor configurations for predicting thermal deformations is presented. From initially 300 temperature sensors, the number of required sensors was significantly reduced while maintaining predictive accuracy. Furthermore, a pattern for sensor placement was identified, providing the potential for an efficient sensor layout that enables cost-effective data acquisition and improved monitoring of machining and wear progression of machine tool components. © 2025 The Authors

关键词： Metal working tools

Exploring the Potential of Virtual Reality for Child Pedestrian Safety Training: A Case Study in China

学校读者我要写书评

暂无评论

Exploring the Potential of Virtual Reality for Child Pedestr...

International Symposium on Educational technology (ISET)

作者： Qinna Feng Gege Li Tingting Yang Heng Luo School of Educational Information Technology Central China Normal University Wuhan China Department of Learning Design and Technology University of Georgia Georgia USA

ISBN: (数字)9781728171890

ISBN: (纸本)9781728171906

This study explores the potential of using virtual reality in conducting child pedestrian safety training by designing and implementing a VR-based training intervention in three elementary schools in China, located in urban, rural, and fringe area respectively. The results suggested enhanced authenticity and interactivity as two technical affordances of VR, which enabled risky behavior diagnosis and personalized remedial instruction. The increased pedestrian performance among children during and after the VR intervention supported its efficacy as an innovative training program. Additionally, the empirical results indicated a limited impact of school location and family SES on children's pedestrian performance, highlighting the need for a universal pedestrian safety training curriculum for elementary schools nationwide.

关键词： Training Safety Green products Virtual environments Interviews