Review
Multidisciplinary Sciences
Zhi-Hua Zhou
Summary: This article introduces the concept and challenges of open-environment machine learning, as well as some technological advances and theoretical issues in this field.
NATIONAL SCIENCE REVIEW
(2022)
Article
Chemistry, Multidisciplinary
Krzysztof Zatwarnicki, Waldemar Pokuta, Anna Bryniarska, Anna Zatwarnicka, Andrzej Metelski, Ewelina Piotrowska
Summary: The article introduces a new language called General Environment Description Language (GEDL) which has the potential to become the basis for building general purpose intelligent systems in the future. The motivation for the research is presented based on previous works in the field. The article provides an overview of the concept of the language and basic definitions of its elements, along with an example of GEDL language usage in JSON notation.
APPLIED SCIENCES-BASEL
(2021)
Article
Computer Science, Artificial Intelligence
Bryan Allen Plummer, Kevin Shih, Yichen Li, Ke Xu, Svetlana Lazebnik, Stan Sclaroff, Kate Saenko
Summary: This paper proposes an approach for associating image regions and phrases by extending Faster R-CNN and initializing classification layers using canonical correlation analysis (CCA), resulting in significant performance improvement on three popular phrase grounding datasets.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Education, Scientific Disciplines
Lotta Coenen, Laura Daman, Guy Gielis, An Stockmans, Arne Van Renterghem, Eline Maes, Marc Van Nuland, Nele R. Michels
Summary: This study developed a 360-degree evaluation tool called TOEKAN to improve the quality of GP training practices. It involves all stakeholders and provides regular surveys with access to the results. By continuously monitoring and improving this evaluation tool, the quality of clinical learning environment in GP education can be enhanced.
Article
Computer Science, Information Systems
Nicolas Loizou, Peter Richtarik
Summary: This work introduces a new framework for analyzing and designing randomized gossip algorithms to solve the average consensus problem, recovering well-known algorithm variants and proposing new specific gossip methods. Through extensive experimental testing, the performance of the proposed gossip protocols is evaluated in typical wireless network topologies.
IEEE TRANSACTIONS ON INFORMATION THEORY
(2021)
Article
Education & Educational Research
Jose Armando Valente, Ricardo Caceffo, Rodrigo Bonacin, Julio Cesar dos Reis, Diego Addan Goncalves, Maria Cecilia Calani Baranauskas
Summary: This study explores how an embodied-based constructionist environment can help kindergarten children interact with digital technologies, finding that children's sequence of performative actions in this environment differs from traditional constructionist environments, emphasizing their ability to solve specific challenges through action and collaboration.
BRITISH JOURNAL OF EDUCATIONAL TECHNOLOGY
(2021)
Article
Engineering, Electrical & Electronic
Xin Li, Weisheng Dong, Jinjian Wu, Leida Li, Guangming Shi
Summary: Superresolution (SR) imaging, a classic problem in multidimensional signal processing, involves reconstructing high-resolution (HR) images from their low-resolution (LR) counterparts. This tutorial article reviews the development of SR technology, focusing on the evolution of key insights from analytical representations to data-driven deep models. Both model-based and learning-based approaches are discussed, including different priors and neural network architectures. The limitations of both approaches are highlighted, leading to a hybrid approach that combines their strengths. Open challenges in SR, such as arbitrary-ratio, reference-based, and domain-specific SR, are also discussed.
IEEE SIGNAL PROCESSING MAGAZINE
(2023)
Review
Computer Science, Artificial Intelligence
Rafael Figueiredo Prudencio, Marcos R. O. A. Maximo, Esther Luna Colombini
Summary: With the widespread adoption of deep learning, reinforcement learning (RL) has become increasingly popular in solving complex problems such as playing complex games, sustaining conversations with humans, and controlling robotic agents. However, there are still domains inaccessible to RL due to the high cost and danger of interacting with the environment. Offline RL, which learns exclusively from static datasets, can extract policies from large and diverse training datasets, making it particularly appealing for real-world applications. This work provides a unifying taxonomy to classify offline RL methods, reviews the latest algorithmic breakthroughs, and summarizes the performance of each method and class of methods on different dataset properties, equipping researchers with the tools to make informed decisions and identify promising research directions.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Social Sciences, Interdisciplinary
Hawraz Qader Hama, Ulker Vanci Osam
Summary: This study explored the perceived impact of an Internet-based instructional learning environment (iBILE) on developing communication, collaboration, and reflection skills in microteaching for preservice English language teachers. The results indicated that the designed iBILE was highly effective in addressing microteaching problems and creating unique opportunities for communication, collaboration, and reflection among the preservice teachers.
Article
Psychology, Multidisciplinary
Stefan Kopp, Nicole Kramer
Summary: The study of human-human communication and the development of computational models for human-agent communication have diverged significantly over the past decade. Despite claims of super-human performance in certain tasks, no system is currently capable of engaging in a coherent conversation with a human. The paper argues for a re-evaluation of the hallmarks of cooperative communication and the core capabilities needed for conversational agents based on research on human-human communication and psychological processes.
FRONTIERS IN PSYCHOLOGY
(2021)
Article
Chemistry, Multidisciplinary
Deborah Brady, Krisztina Voronova
Summary: This paper discusses the implementation of a flipped active learning classroom model in large general chemistry classes at the University of Nevada, Reno. The results of a survey on students' perception towards the course design are provided. The effectiveness of different learning environments, including traditional in-person, emergency remote teaching, online, and in-person flipped learning, was evaluated over six semesters. The study highlights the importance of flexibility, organization, clear expectations, and fast communication in course design.
JOURNAL OF CHEMICAL EDUCATION
(2023)
Article
Construction & Building Technology
Anna Kristin Siguroardottir, Torfi Hjartarson, Aoalsteinn Snorrason
Summary: This paper presents a post-occupancy evaluation of a school building in Iceland that combines open and confined spaces for various pedagogical approaches. The study utilized research interviews and pedagogical walk-throughs to assess the strengths and weaknesses of the design, offering valuable insights for those involved in educational space design and application.
Review
Multidisciplinary Sciences
Gergely Mark Csanyi, Daniel Nagy, Renato Vagi, Janos Pal Vadasz, Tamas Orosz
Summary: Data sharing in judicial systems can enhance transparency, but the sensitive information in legal documents must be anonymized to prevent privacy breaches. Named Entity Recognition methods, machine learning, and anonymization models like differential privacy are crucial for reducing re-identification risk.Preserving the utility of the text while anonymizing legal documents is essential.
Article
Engineering, Civil
Shiying Zhang, Jun Li, Long Shi, Ming Ding, Dinh C. Nguyen, Wuzheng Tan, Jian Weng, Zhu Han
Summary: Intelligent transportation systems (ITSs) face challenges in making timely and accurate decisions of vehicle behaviors due to the dynamic characteristics of vehicle networks. Federated learning (FL), a distributed machine learning technology, has gained attention for its privacy protection and scalability. We conducted a comprehensive survey on the developments of FL in ITS, addressing challenges, deployments, potential issues, and analyzing new challenges and limitations introduced by FL, along with collaborative technologies and future research directions.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Computer Science, Interdisciplinary Applications
Xiaohan Wang, Lin Zhang, Tingyu Lin, Chun Zhao, Kunyu Wang, Zhen Chen
Summary: This paper proposes a multi-agent reinforcement learning algorithm to solve job scheduling problems in a resource preemption environment. By modeling the resource preemption environment as a decentralized partially observable Markov decision process and constructing a multi-agent scheduling architecture, the decision-making policy of each agent and the cooperation between job agents are learned. The experimental results demonstrate the superiority of the proposed method in terms of total makespan, training stability, and model generalization, compared to traditional rule-based methods and distributed-agent reinforcement learning methods.
ROBOTICS AND COMPUTER-INTEGRATED MANUFACTURING
(2022)
Article
Multidisciplinary Sciences
Marc G. Bellemare, Salvatore Candido, Pablo Samuel Castro, Jun Gong, Marlos C. Machado, Subhodeep Moitra, Sameera S. Ponda, Ziyu Wang
Proceedings Paper
Computer Science, Artificial Intelligence
Marlos C. Machado, Marc G. Bellemare, Michael Bowling
THIRTY-FOURTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, THE THIRTY-SECOND INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE AND THE TENTH AAAI SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE
(2020)
Proceedings Paper
Computer Science, Artificial Intelligence
Craig Sherstan, Adam White, Marlos C. Machado, Patrick M. Pilarski
ARTIFICIAL GENERAL INTELLIGENCE (AGI 2016)
(2016)
Article
Computer Science, Interdisciplinary Applications
Renato Luiz de Freitas Cunha, Marlos C. Machado, Luiz Chaimowicz
COMPUTERS IN ENTERTAINMENT
(2014)
Article
Computer Science, Artificial Intelligence
Erik Talvitie, Satinder Singh
JOURNAL OF ARTIFICIAL INTELLIGENCE RESEARCH
(2011)