Article
Computer Science, Artificial Intelligence
Dong Li, Shulin Liu, Furong Gao, Xin Sun
Summary: C-CLCM is a continual learning classification method inspired by the biological immune system. It gradually enhances its performance by continually learning new types of data during the testing stage. When degenerating into a common supervised learning classification method, it outperforms other methods, especially when the training data do not cover all types.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Automation & Control Systems
Altaf Hussain, Samee Ullah Khan, Noman Khan, Mohammad Shabaz, Sung Wook Baik
Summary: The integration of artificial intelligence (AI) into human activity recognition (HAR) in smart surveillance systems has the potential to revolutionize behavior monitoring, improving security and surveillance measures. A proposed AI-based behavior biometrics framework is introduced, utilizing a dynamic attention fusion unit (DAFU) and temporal-spatial fusion (TSF) network to effectively recognize human activity in surveillance systems.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2024)
Article
Physics, Multidisciplinary
Agnieszka Duraj, Daniel Duczyminski
Summary: The present article aims to find the most efficient machine learning method for detecting abnormal segments in human movement. A new method based on a nested binary classifier is proposed and compared with deep neural networks. Test experiments using popular machine learning algorithms show that the method of nested binary classifiers is an effective way to recognize outlier patterns for HAR systems.
Article
Computer Science, Information Systems
Emad Mabrouk, Yara Raslan, Abdel-Rahman Hedar
Summary: This paper introduces a machine learning method based on the metaheuristics programming framework and immune system programming algorithm, aiming to solve the disruption problem of basic operations in traditional genetic programming methods. The efficiency of the proposed method is validated through numerical experiments.
Article
Computer Science, Artificial Intelligence
Fareeha Rasheed, Abdul Wahid
Summary: This study examines the importance of learning styles in helping students retain concepts and improve understanding, using questionnaires and attribute identification to identify learner styles. Classification algorithms were implemented to analyze the data set, revealing interesting patterns in learner behavior across different contexts and concepts.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Chemistry, Multidisciplinary
Michael B. Bolger
Summary: This perspective article highlights the Chemistry Classification System (CCS) as a flexible and innovative approach to drug classification, emphasizing the use of machine-learning models and AI to guide formulation development based on unique physicochemical properties.
JOURNAL OF CONTROLLED RELEASE
(2022)
Article
Computer Science, Theory & Methods
Jose Mena, Oriol Pujol, Jordi Vitria
Summary: This study introduces the importance of uncertainty estimation in machine learning systems and analyzes how uncertainty can be measured in classification systems based on deep learning. The study also provides an overview of practical considerations in different applications and highlights the properties that should be considered when developing metrics.
ACM COMPUTING SURVEYS
(2022)
Article
Computer Science, Artificial Intelligence
Dong Li, Lanlan Gong, Shulin Liu, Xin Sun, Ming Gu, Kun Qian
Summary: This paper proposes a classification method based on the continual learning mechanism of the biological immune system. It uses single-label memory cells to identify the types of testing data and continuously learns new data to enhance its classification ability. Experimental results show that the method performs well as both a standard batch learning classification method and in handling unseen types of data.
JOURNAL OF INTELLIGENT & FUZZY SYSTEMS
(2022)
Article
Computer Science, Information Systems
Galina Kamyshova, Aleksey Osipov, Sergey Gataullin, Sergey Korchagin, Stefan Ignar, Timur Gataullin, Nadezhda Terekhova, Stanislav Suvorov
Summary: This article proposes a methodology for optimizing crop irrigation using a phytoindication system based on computer vision methods. The system, divided into three stages including image preprocessing, classification, and neural network training, achieves a high accuracy rate of 93% in plant identification. The system can process up to 100 plants per second, surpassing similar systems in performance.
Article
Environmental Sciences
Takuya Kurihana, Elisabeth J. Moyer, Ian T. Foster
Summary: This study introduces a new analysis approach that uses artificial intelligence to classify satellite cloud observations, reducing the dimensionality of the data and generating a unique cloud dataset. The method captures a greater variety of cloud types and provides rich information for global analysis, contributing to the advancement of climate research.
Article
Computer Science, Artificial Intelligence
Marco Alfonse, Mariam Gawich
Summary: The study focuses on automated news classification and presents a novel methodology for classifying Arabic news with an accuracy of over 95%. Researchers often prioritize English news classification and overlook Arabic news due to its complex morphology.
WILEY INTERDISCIPLINARY REVIEWS-DATA MINING AND KNOWLEDGE DISCOVERY
(2022)
Review
Fisheries
Juan Li, Wenkai Xu, Limiao Deng, Ying Xiao, Zhongzhi Han, Haiyong Zheng
Summary: This article reviews the current application status, challenges, and future directions of DL in the field of aquatic animal recognition and detection. Key advances and applications of DL in the recognition and detection of aquatic animals are summarized. DL has significant application value in the field of aquatic animal recognition and detection.
REVIEWS IN AQUACULTURE
(2023)
Article
Chemistry, Analytical
Lukas Picek, Milan Sulc, Jiri Matas, Jacob Heilmann-Clausen, Thomas S. Jeppesen, Emil Lind
Summary: This article presents an AI-based fungi species recognition system that collaborates with a citizen-science community, resulting in real-time identification, increased data collection, and improved classification using a novel method based on a Vision Transformer architecture.
Article
Physics, Multidisciplinary
Sandra Castellanos-Paez, Nicolas Hili, Alexandre Albore, Mar Perez-Sanagustin
Summary: Paper and pens are commonly used tools by systems engineers, but digitizing sketched models into computer tools remains difficult and error-prone. Machine learning methods can improve sketch recognition performance, but lack of explainability limits trust. This study proposes an approach combining symbolic AI and machine learning to recognize text and symbols separately, improving performance while preserving explainability.
FRONTIERS IN PHYSICS
(2022)
Article
Energy & Fuels
Paria Movahed, Saman Taheri, Ali Razban
Summary: Long-term operation of HVAC systems can lead to failures, higher energy consumption, and maintenance costs. Fault detection diagnostic (FDD) is commonly used to prevent malfunctions, and machine learning methods have gained interest due to their high accuracy. However, existing studies suffer from biased classification algorithms and high false positives. To address these challenges and improve diagnostic performance, this study proposes a novel data-driven framework using principal component analysis, time series anomaly detection, and random forest.
