Article
Computer Science, Artificial Intelligence
Chunnan Wang, Bozhou Chen, Geng Li, Hongzhi Wang
Summary: This paper proposes an efficient GNN NAS algorithm called FL-AGNNS, which enables distributed agents to cooperatively design powerful GNN models while preserving personal information. FL-AGNNS uses a federated evolutionary optimization strategy and weight sharing strategy to recommend GNN architectures that perform well on multiple datasets.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Geochemistry & Geophysics
Cheng Peng, Yangyang Li, Licheng Jiao, Ronghua Shang
Summary: This article investigates a new paradigm to automatically design a suitable CNN architecture for scene classification, achieving impressive classification performance on remote sensing scene images through an efficient architecture search framework and dataset merging strategy.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2021)
Article
Geochemistry & Geophysics
Jie Chen, Haozhe Huang, Jian Peng, Jiawei Zhu, Li Chen, Chao Tao, Haifeng Li
Summary: This article proposes a contextual information-preserved architecture learning framework for remote-sensing scene classification. By incorporating channel compression and adding potential operators, the method reduces time consumption and discovers new architectures that better utilize contextual information in remote-sensing scenes.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Computer Science, Artificial Intelligence
Hugo Touvron, Piotr Bojanowski, Mathilde Caron, Matthieu Cord, Alaaeldin El-Nouby, Edouard Grave, Gautier Izacard, Armand Joulin, Gabriel Synnaeve, Jakob Verbeek, Herve Jegou
Summary: ResMLP is an image classification architecture that relies solely on multi-layer perceptrons. It achieves accurate and efficient results through a combination of linear layers and a two-layer feed-forward network, trained with modern strategies and data augmentation.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Engineering, Electrical & Electronic
Yin Tang, Qi Teng, Lei Zhang, Fuhong Min, Jun He
Summary: The research introduces a lightweight CNN model using Lego filters for HAR tasks, achieving higher accuracy and reducing memory and computation cost. Experimental results show that this Lego CNN model is smaller, faster, and more accurate than traditional CNN. Moreover, the model does not rely on special network structures and has application potential in ubiquitous and wearable computing.
IEEE SENSORS JOURNAL
(2021)
Article
Computer Science, Artificial Intelligence
Xiangru Chen, Yanan Sun, Mengjie Zhang, Dezhong Peng
Summary: In this article, a novel method for automatically designing optimal architectures of VAEs for image classification, called EvoVAE, based on a genetic algorithm, is proposed. Experiment results demonstrate the superiority of the EvoVAE algorithm over peer competitors on three benchmark datasets.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2021)
Article
Computer Science, Information Systems
Babatounde Moctard Oloulade, Jianliang Gao, Jiamin Chen, Tengfei Lyu, Raeed Al-Sabri
Summary: This paper provides a comprehensive review of automatic GNN model building frameworks to facilitate future progress. The complexity of GNN model components poses challenges to existing GNN efficiency, leading to the development of automated machine learning frameworks for finding the best GNN models.
TSINGHUA SCIENCE AND TECHNOLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Noha W. Hasan, Ali S. Saudi, Mahmoud I. Khalil, Hazem M. Abbas
Summary: In this study, an efficient genetic algorithm is proposed to find optimized CNN architectures for the acoustic scene classification task. The algorithm utilizes frequency-dimension splitting and explores the architecture search space of sub-CNN models to better capture the distinct features of the input spectrograms.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2023)
Article
Computer Science, Artificial Intelligence
Yunho Jeon, Junmo Kim
Summary: This study focuses on the convolution units within convolutional networks, proposing an active convolution unit (ACU) and grouped ACU, with a detailed analysis showing their efficiency. Experimental results demonstrate that these proposed units can replace existing convolution structures.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2021)
Article
Computer Science, Software Engineering
Bowen Xu, Thong Hoang, Abhishek Sharma, Chengran Yang, Xin Xia, David Lo
Summary: This study proposes a specialized deep learning architecture, Post2Vec, for analyzing Stack Overflow posts and extracting distributed representations. The effectiveness of Post2Vec is demonstrated through its application in tag recommendation and other tasks.
IEEE TRANSACTIONS ON SOFTWARE ENGINEERING
(2022)
Article
Computer Science, Artificial Intelligence
Antonio-Javier Gallego, Jorge Calvo-Zaragoza, Robert B. Fisher
Summary: In the context of supervised statistical learning, the study explored the issue of inconsistent distributions between training and test sets, presenting an incremental approach to address it. By utilizing an unsupervised domain adaptation algorithm to identify target samples and iteratively adapting the model through self-labeling, an adversarial training strategy was proposed to enhance the performance of domain adaptation algorithms.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Qiang Gao, Zhipeng Luo, Diego Klabjan, Fengli Zhang
Summary: Continual learning with efficient architecture search (CLEAS) is proposed to overcome the challenges of catastrophic forgetting, adapting to new tasks, and controlling model complexity. By leveraging neuron-level NAS, CLEAS achieves higher classification accuracy by reusing old neurons and adding new ones on simpler neural architectures.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Engineering, Civil
Chengju Zhou, Meiqing Wu, Siew-Kei Lam
Summary: In this paper, a unified multi-task learning architecture for fast and accurate pedestrian detection is proposed. By integrating a lightweight semantic segmentation branch and optimized modules, the architecture effectively combines pedestrian detection and semantic segmentation tasks, achieving improved detection performance without increasing computational overhead and achieving high detection performance with low resolution input images.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2022)
Article
Computer Science, Theory & Methods
Nan Wu, Yuan Xie
Summary: This article explores the application of machine learning in computer architecture and system design, presenting a comprehensive review of the work in this field. It discusses the role of machine learning techniques in architecture/system design and summarizes the problems that can be solved using these techniques. The article also provides insights into the potential directions and opportunities for applying machine learning in this domain.
ACM COMPUTING SURVEYS
(2023)
Article
Computer Science, Artificial Intelligence
Timothy Hospedales, Antreas Antoniou, Paul Micaelli, Amos Storkey
Summary: The field of meta-learning, or learning-to-learn, has gained significant interest in recent years. Unlike conventional approaches to AI, meta-learning aims to improve the learning algorithm itself by utilizing multiple learning experiences. This provides an opportunity to address challenges in deep learning, including data and computation limitations, as well as generalization.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Multidisciplinary Sciences
Manas Gaur, Vamsi Aribandi, Amanuel Alambo, Ugur Kursuncu, Krishnaprasad Thirunarayan, Jonathan Beich, Jyotishman Pathak, Amit Sheth
Summary: The use of deep learning algorithms to assess suicide risk based on Reddit data shows potential in improving suicide risk assessment, particularly when integrated with clinical diagnostic interviews for greater accuracy. The time-variant approach outperforms the time-invariant method in assessing suicide-related ideations and behaviors, while the time-invariant model performs better in predicting suicide-related behaviors and attempts.
Article
Chemistry, Analytical
Kaishu Xia, Clint Saidy, Max Kirkpatrick, Noble Anumbe, Amit Sheth, Ramy Harik
Summary: The manufacturing industry is currently undergoing a paradigm shift from traditional control pyramids to decentralized, service-oriented, and cyber-physical systems in the Fourth Industrial Revolution. The authors aim to develop a novel CPS-enabled control architecture that includes intelligent information systems, fast and secure industrial communication networks, cognitive automation, and interoperability between machines and humans.
Article
Computer Science, Software Engineering
Utkarshani Jaimini, Amit Sheth
Summary: Humans utilize causality and hypothetical retrospection in decision-making process, planning, and comprehension of life events. The human mind possesses an inherent understanding of causality, developing a model of the world that learns with minimal data points and contemplates alternative scenarios.
IEEE INTERNET COMPUTING
(2022)
Article
Engineering, Electrical & Electronic
Sandeep K. Chaudhuri, Joshua W. Kleppinger, OmerFaruk Karadavut, Ritwik Nag, Rojina Panta, Forest Agostinelli, Amit Sheth, Utpal N. Roy, Ralph B. James, Krishna C. Mandal
Summary: In this article, the growth of Cd0.9Zn0.1Te0.97Se0.03 (CZTS) wide bandgap semiconductor single crystals for room temperature gamma-ray detection is reported. The article also introduces a deep convolutional neural network (CNN) that efficiently identifies the energy of gamma photons detected by a CZTS detector. The CNN is trained using simulated data and shows low prediction error for different energy levels.
JOURNAL OF MATERIALS SCIENCE-MATERIALS IN ELECTRONICS
(2022)
Article
Computer Science, Software Engineering
Manas Gaur, Kalpa Gunaratna, Shreyansh Bhatt, Amit Sheth
Summary: Knowledge-infused learning combines knowledge with data-driven deep learning techniques, improving performance, interpretability, and control over AI systems. This approach introduces symbolic AI into data-driven AI, creating a class of neuro-symbolic AI methods.
IEEE INTERNET COMPUTING
(2022)
Article
Computer Science, Artificial Intelligence
Ruwan Wickramarachchi, Cory Henson, Amit Sheth
Summary: Knowledge-based entity prediction (KEP) is a task that aims to improve machine perception in autonomous systems by predicting potentially unrecognized entities using relational knowledge. This article provides a formal definition of KEP and introduces three potential solutions. The applicability of KEP is demonstrated through its use in autonomous driving and smart manufacturing. The article argues that using KEP in complex real-world systems can significantly enhance machine perception and move current technology closer to achieving full autonomy.
IEEE INTELLIGENT SYSTEMS
(2022)
Editorial Material
Computer Science, Software Engineering
Amit Sheth, Manas Gaur, Kaushik Roy, Revathy Venkataraman, Vedant Khandelwal
Summary: The adoption of AI in high-value, sensitive, or safety-critical applications has been challenging. Incorporating process knowledge (PK) along with domain knowledge in machine learning can improve performance and user-level explainability. This article discusses the importance of PK and presents methods for infusing PK into AI systems.
IEEE INTERNET COMPUTING
(2022)
Article
Computer Science, Software Engineering
Utkarshani Jaimini, Tongtao Zhang, Georgia Olympia Brikis, Amit Sheth
Summary: The Metaverse, a hyperrealistic virtual world, holds great potential in the future of the Internet. Predictions indicate that people will spend more time in the Metaverse and organizations will develop products and services for this virtual realm.
IEEE INTERNET COMPUTING
(2022)
Article
Public, Environmental & Occupational Health
Usha Lokala, Francois Lamy, Raminta Daniulaityte, Manas Gaur, Amelie Gyrard, Krishnaprasad Thirunarayan, Ugur Kursuncu, Amit Sheth
Summary: This article describes the development and application of the drug abuse ontology (DAO) for analyzing web-based and social media data in substance use epidemiology research. The DAO provides a flexible framework and a useful resource for big data analytics.
JMIR PUBLIC HEALTH AND SURVEILLANCE
(2022)
Article
Computer Science, Artificial Intelligence
Amit Sheth, Kaushik Roy, Manas Gaur
Summary: Humans interact with the environment by transforming perception into symbols and mapping symbols to knowledge for abstraction, reasoning, and planning. Machine perception in AI refers to pattern recognition from data trained using self-supervised learning, while machine cognition involves more complex computations based on knowledge of the environment. Humans can control and explain their cognitive functions, requiring retention of symbolic mappings from perception to environment knowledge. This is important in safety-critical applications like healthcare, criminal justice, and autonomous driving.
IEEE INTELLIGENT SYSTEMS
(2023)
Article
Biotechnology & Applied Microbiology
Deepa Tilwani, Jessica Bradshaw, Amit Sheth, Christian O'Reilly
Summary: In recent years, there has been an increasing prevalence of autism spectrum disorder (ASD), but diagnosis typically occurs later than the optimal intervention age. This study investigated the potential of electrocardiogram (ECG) recordings as an ASD biomarker in infants aged 3-6 months. The findings suggest that ECG signals contain important information about ASD likelihood and the KNN machine learning algorithm performed best in classifying ASD likelihood.
BIOENGINEERING-BASEL
(2023)
Article
Health Care Sciences & Services
Khushi S. Patel, Cynthia F. Corbett, Elizabeth M. Combs, Sara B. Donevant, Margaret J. Selph, Lynette M. Gibson, Robin M. Dawson, Amit P. Sheth, Ronda G. Hughes
Summary: This study aimed to explore the perceptions and experiences of uninsured individuals seeking care at a free medical clinic regarding COVID-19, as well as their suggestions for adapting a mobile health app for this population. The findings from a survey of 240 respondents and interviews with a subset of participants revealed that most respondents had limited awareness of the risk of COVID-19, but they expressed interest in using a mobile health app to gain more information.
JMIR FORMATIVE RESEARCH
(2023)
Article
Computer Science, Artificial Intelligence
Fadi El Kalach, Ruwan Wickramarachchi, Ramy Harik, Amit Sheth
Summary: The next phase of manufacturing involves transitioning from traditional automated systems to autonomous systems that can reallocate resources as needed, minimize downtime, and meet market demands. This article discusses the progress made in applying Semantic Web capabilities to manage production lines and showcases a fully autonomous manufacturing use case that integrates diverse data sources and domain knowledge.
IEEE INTELLIGENT SYSTEMS
(2023)
Article
Computer Science, Software Engineering
Revathy Venkataramanan, Aalap Tripathy, Martin Foltin, Hong Yung Yip, Annmary Justine, Amit Sheth
Summary: AI pipelines are complex and expensive to execute. By analyzing metadata of past executed pipelines, relevant parameters can be identified and recommended to users, improving the optimization of AI experiments.
IEEE INTERNET COMPUTING
(2023)
Article
Computer Science, Software Engineering
Manas Gaur, Keyur Faldu, Amit Sheth
Summary: The recent innovations in deep learning have shown great potential but also brought challenges, including interpretability and utilization of knowledge. Knowledge-infused learning (K-iL) provides a solution by incorporating knowledge graphs.
IEEE INTERNET COMPUTING
(2021)
