Review
Computer Science, Information Systems
Raymon van Dinter, Bedir Tekinerdogan, Cagatay Catal
Summary: This study conducts a systematic literature review on the automation of SLR studies, aiming to collect and synthesize the current research in this area for further exploration. The review analyzes 41 primary studies and identifies the objectives, application domains, automated steps, techniques, challenges, and solution directions in SLR automation. The study reveals that while automation approaches for SLR have focused on the selection of primary studies, there is a lack of automation techniques applied in the planning and reporting phases, indicating a need for further research in automating other activities of the SLR process.
INFORMATION AND SOFTWARE TECHNOLOGY
(2021)
Review
Medicine, General & Internal
Kevin E. K. Chai, Robin L. J. Lines, Daniel F. Gucciardi, Leo Ng
Summary: The use of Research Screener, a semi-automated machine learning tool, has been shown to significantly reduce the burden of screening titles and abstracts in systematic reviews, saving time and cost. Through simulation and real-world examples, the tool can provide workload savings of 60% to 96%, as well as substantial time and financial benefits.
SYSTEMATIC REVIEWS
(2021)
Article
Computer Science, Artificial Intelligence
Rens van de Schoot, Jonathan de Bruin, Raoul Schram, Parisa Zahedi, Jan de Boer, Felix Weijdema, Bianca Kramer, Martijn Huijts, Maarten Hoogerwerf, Gerbrich Ferdinands, Albert Harkema, Joukje Willemsen, Yongchao Ma, Qixiang Fang, Sybren Hindriks, Lars Tummers, Daniel L. Oberski
Summary: The new open source machine learning framework ASReview, utilizing active learning and a variety of machine learning models, can efficiently and systematically check the literature for systematic reviews.
NATURE MACHINE INTELLIGENCE
(2021)
Article
Health Care Sciences & Services
Xuan Qin, Jiali Liu, Yuning Wang, Yanmei Liu, Ke Deng, Yu Ma, Kang Zou, Ling Li, Xin Sun
Summary: This study examines the effectiveness of NLP technology in assisting rapid literature screening when updating systematic reviews. Using a LightGBM model, the study achieved high sensitivity and specificity, indicating that NLP technology can reduce reviewers' workload and improve screening efficiency.
JOURNAL OF CLINICAL EPIDEMIOLOGY
(2021)
Article
Green & Sustainable Science & Technology
Nicla Frigerio, Claudio F. A. Cornaggia, Andrea Matta
Summary: This study explores energy efficient control measures for machining processes by utilizing the machine power state control. By extending control policies through online learning and real-time data acquisition, a novel optimization algorithm is proposed to minimize energy consumption and ensure target production rates.
JOURNAL OF CLEANER PRODUCTION
(2021)
Review
Biology
Kaitlyn Hair, Zsanett Bahor, Malcolm Macleod, Jing Liao, Emily S. Sena
Summary: This study presents the performance of ASySD, an automated tool for deduplicating systematic searches in biomedical reviews. ASySD outperforms both SRA-DM and EndNote in identifying and removing duplicates, with a sensitivity of 0.95 to 0.99 and a specificity of > 0.99. The tool is time-saving, reliable, and freely available online.
Article
Biology
Jonathan Bouvette, Qinwen Huang, Amanda A. Riccio, William C. Copeland, Alberto Bartesaghi, Mario J. Borgnia
Summary: This article introduces SmartScope, a framework for streamlining, standardizing, and automating the evaluation of cryo-EM specimens. Through deep-learning-based object detection, SmartScope can automatically identify and classify features suitable for imaging, enabling fully automated specimen screening. It provides a web interface for remote real-time control of microscope operation and access to images and annotation tools. Manual annotations can be used to re-train the feature recognition models for improved performance. This automated tool simplifies structure determination and lowers the barrier of adoption for cryo-EM.
Review
Mathematical & Computational Biology
Ashley Elizabeth Muller, Heather Melanie R. Ames, Patricia Sofia Jacobsen Jardim, Christopher James Rose
Summary: The study evaluated the utility of an automated clustering method in categorizing studies, finding that automated clustering had higher precision and recall compared to manual categorization, and also saved 49% more time. The clustering algorithm was sensitive enough to group studies based on linguistic differences, corresponding to manual categories.
RESEARCH SYNTHESIS METHODS
(2022)
Article
Medicine, General & Internal
John Zimmerman, Robin E. Soler, James Lavinder, Sarah Murphy, Charisma Atkins, LaShonda Hulbert, Richard Lusk, Boon Peng Ng
Summary: The article discusses an approach that utilizes existing examples of SRs to develop and test a method for assisting the SR title and abstract pre-screening by reducing the initial pool of potential articles down to those that meet inclusion criteria. This approach differs from previous uses of ML in SR in that it incorporates ML configurations guided by previously conducted SRs and human confirmation on ML predictions of relevant articles during iterative reviews. The study shows promising results in achieving high sensitivity in finding relevant articles while reducing the human review workload.
SYSTEMATIC REVIEWS
(2021)
Review
Medicine, Research & Experimental
Antti A. A. Maekitie, Rasheed Omobolaji Alabi, Sweet Ping Ng, Robert P. P. Takes, K. Thomas Robbins, Ohad Ronen, Ashok R. R. Shaha, Patrick J. J. Bradley, Nabil F. F. Saba, Sandra Nuyts, Asterios Triantafyllou, Cesare Piazza, Alessandra Rinaldo, Alfio Ferlito
Summary: This article provides an analysis of systematic reviews on the current status and limitations of the application of artificial intelligence (AI) and machine learning (ML) as decision-making tools in head and neck cancer (HNC) management. The reviews reveal that AI/ML can be used for detecting cancerous lesions, predicting histopathological nature, prognosticating, extracting pathological findings from imaging, and various applications in radiation oncology. However, the lack of standardized guidelines, performance reporting, external validation procedures, and regulatory frameworks limit their adoption in clinical practice.
ADVANCES IN THERAPY
(2023)
Article
Health Care Sciences & Services
Patricia Sofia Jacobsen Jardim, Christopher James Rose, Heather Melanie Ames, Jose Francisco Meneses Echavez, Stijn Van de Velde, Ashley Elizabeth Muller
Summary: This study aimed to assess the feasibility of using RobotReviewer for automated risk of bias assessment, and the results showed that it was as accurate as human assessment, but there were differences in acceptability among researchers. Some less experienced reviewers were positive towards the tool, while others emphasized the importance of human input and interaction.
BMC MEDICAL RESEARCH METHODOLOGY
(2022)
Article
Mathematical & Computational Biology
Allard J. van Altena, Rene Spijker, Mariska M. G. Leeflang, Silvia Delgado Olabarriaga
Summary: This paper introduces an approach to select data for model training and compares it with established methods using 50 Cochrane diagnostic test accuracy reviews. The study suggests that models perform best with a larger number of reviews in the training set and when the research subject of the target review is similar to other reviews in the dataset.
RESEARCH SYNTHESIS METHODS
(2021)
Article
Social Issues
Willian Boschetti Adamczyk, Leonardo Monasterio, Adelar Fochezatto
Summary: This study explores the impact of automation on public sector employment in Brazil, estimating that approximately 20% of public sector employees are in jobs with a high potential for automation. Government occupations with lower education and salary levels are most susceptible to future automation.
TECHNOLOGY IN SOCIETY
(2021)
Article
Construction & Building Technology
Pablo Martinez, Beda Barkokebas, Farook Hamzeh, Mohamed Al-Hussein, Rafiq Ahmad
Summary: Offsite construction is a method focused on improving productivity by moving construction tasks to manufacturing facilities. This paper presents a novel approach that combines deep learning algorithms and finite state machines to automatically detect and track the progress of construction operations.
AUTOMATION IN CONSTRUCTION
(2021)
Article
Mathematical & Computational Biology
Toni Lange, Guido Schwarzer, Thomas Datzmann, Harald Binder
Summary: The research project explored the potential of machine learning methods to reduce human workload, evaluated the performance of deep learning methods compared to more established machine learning methods, and highlighted the importance of data preprocessing on the final performance of approaches. Machine learning methods provided reasonable classification, but the final performance heavily relies on data preparation.
RESEARCH SYNTHESIS METHODS
(2021)