Article
Computer Science, Hardware & Architecture
Isabelle Hupont, Marina Micheli, Blagoj Delipetrev, Emilia Gomez, Josep Soler Garrido
Summary: This article discusses transparency obligations in the proposed European regulatory framework for artificial intelligence (AI), known as the Artificial Intelligence Act. It analyzes the effectiveness of current approaches for AI documentation in meeting requirements and assesses their suitability as a foundation for future technical standards.
Editorial Material
Biochemistry & Molecular Biology
Eric Wu, Kevin Wu, Roxana Daneshjou, David Ouyang, Daniel E. Ho, James Zou
Summary: A comprehensive overview of medical AI devices approved by the US Food and Drug Administration sheds light on limitations of the evaluation process that may mask vulnerabilities of devices when deployed on patients.
Article
Information Science & Library Science
Crispin Coombs, Patrick Stacey, Peter Kawalek, Boyka Simeonova, Joerg Becker, Katrin Bergener, Joao Alvaro Carvalho, Marcelo Fantinato, Niels F. Garmann-Johnsen, Christian Grimme, Armin Stein, Heike Trautmann
Summary: Researchers discussed the relationship between artificial intelligence and humanity, highlighting the importance of preserving human attributes such as criticism of AI design and application, human involvement in intelligent machine decision-making processes, and the ability to interpret intelligent machine decision-making processes.
INTERNATIONAL JOURNAL OF INFORMATION MANAGEMENT
(2021)
Editorial Material
Health Care Sciences & Services
Mirja Mittermaier, Marium M. Raza, Joseph C. Kvedar
Summary: Artificial intelligence is increasingly used in healthcare, particularly in surgery. While it holds promise in predicting outcomes and guiding surgeons, AI systems can also be biased, exacerbating existing inequalities. This impacts disadvantaged populations, who may receive less accurate algorithmic predictions or underestimate their need for care. Detecting and mitigating bias is crucial for creating fair and generalizable AI technology. This article discusses a recent study that developed a new strategy to address bias in surgical AI systems.
NPJ DIGITAL MEDICINE
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Jonas Andersson, Tufve Nyholm, Crister Ceberg, Anja Almen, Peter Bernhardt, Annette Fransson, Lars E. Olsson
Summary: The survey of Swedish medical physicists reveals a positive attitude towards AI, but a general low self-assessment of knowledge and preparedness for AI. The study concludes that AI will change the medical physics profession, providing opportunities but also highlighting the need for increased education and training on AI to overcome potential weaknesses.
PHYSICA MEDICA-EUROPEAN JOURNAL OF MEDICAL PHYSICS
(2021)
Article
Multidisciplinary Sciences
Vajira Thambawita, Pegah Salehi, Sajad Amouei Sheshkal, Steven A. Hicks, Hugo L. Hammer, Sravanthi Parasa, Thomas de Lange, Pal Halvorsen, Michael A. Riegler
Summary: The study introduces a novel synthetic data generation pipeline SinGAN-Seg to produce synthetic medical images with corresponding masks using a single training image. The pipeline significantly improves the quality of generated data and enhances the performance of segmentation models when training datasets do not have a considerable amount of data.
Article
Radiology, Nuclear Medicine & Medical Imaging
Summary: A survey conducted among European Society of Radiology members examines the practical clinical experience of radiologists with AI-powered tools. The majority of radiologists experienced no reduction in workload, but AI algorithms were found to be reliable for different use case scenarios.
INSIGHTS INTO IMAGING
(2022)
Article
Health Care Sciences & Services
Emilia Niemiec
Summary: This article discusses the impact of the Medical Device Regulation on improving the safety and performance of medical artificial intelligence devices. The Regulation introduces new requirements for risk classification, clinical evaluation, post-market surveillance, and notified bodies, which aim to enhance the safety and performance of these devices. The guidance provided by the Medical Device Coordination Group also addresses some of the issues identified in studies on medical artificial intelligence devices.
Review
Gastroenterology & Hepatology
Cameron Stewart, Stephen K. Y. Wong, Joseph J. Y. Sung
Summary: The development of AI and digital health in healthcare brings up concerns regarding equitable access, data privacy, inclusiveness, bias, discrimination, and the clinician-patient relationship. This article discusses ethical and legal issues in using AI in gastroenterology, emphasizing principles such as respect for individuals, privacy, integrity, conflict of interest, beneficence, nonmaleficence, and justice. While a principle-based approach is currently recommended for problem-solving, future efforts should focus on addressing more specific and concrete issues related to AI in healthcare.
JOURNAL OF GASTROENTEROLOGY AND HEPATOLOGY
(2021)
Article
Multidisciplinary Sciences
Maria Agustina Ricci Lara, Rodrigo Echeveste, Enzo Ferrante
Summary: AI systems in the field of medical imaging can exhibit unfair biases, and it is important to address the meaning of fairness and potential sources of biases, as well as implement strategies to mitigate them. An analysis of the current state of the field reveals strengths and areas for improvement, along with challenges and opportunities.
NATURE COMMUNICATIONS
(2022)
Article
Health Care Sciences & Services
Mingxuan Liu, Yilin Ning, Salinelat Teixayavong, Mayli Mertens, Jie Xu, Daniel Shu Wei Ting, Lionel Tim-Ee Cheng, Jasmine Chiat Ling Ong, Zhen Ling Teo, Ting Fang Tan, Narrendar Ravichandran, Fei Wang, Leo Anthony Celi, Marcus Eng Hock Ong, Nan Liu
Summary: Artificial intelligence has shown its ability to extract insights from data, but ensuring fairness in high-stakes fields such as healthcare remains a concern. The notion of fairness in clinical contexts requires careful examination and alignment with ethical considerations. A multidisciplinary approach involving AI researchers, clinicians, and ethicists is necessary to bridge the gap between technical developments and clinical needs.
NPJ DIGITAL MEDICINE
(2023)
Review
Psychology, Multidisciplinary
Tiago Buatim Nion Da Silveira, Heitor Silverio Lopes
Summary: This paper addresses the divergences and contradictions in the definition of intelligence across different areas of knowledge, particularly in computational intelligence and psychology. The lack of a standardized definition, empirical evidence, or measurement strategy for intelligence hinders cross-fertilization between these areas, especially in semantic-based applications.
FRONTIERS IN PSYCHOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Alexandros Karargyris, Renato Umeton, Micah J. Sheller, Alejandro Aristizabal, Johnu George, Anna Wuest, Sarthak Pati, Hasan Kassem, Maximilian Zenk, Ujjwal Baid, Prakash Narayana Moorthy, Alexander Chowdhury, Junyi Guo, Sahil Nalawade, Jacob Rosenthal, David Kanter, Maria Xenochristou, Daniel J. Beutel, Verena Chung, Timothy Bergquist, James Eddy, Abubakar Abid, Lewis Tunstall, Omar Sanseviero, Dimitrios Dimitriadis, Yiming Qian, Xinxing Xu, Yong Liu, Rick Siow Mong Goh, Srini Bala, Victor Bittorf, Sreekar Reddy Puchala, Biagio Ricciuti, Soujanya Samineni, Eshna Sengupta, Akshay Chaudhari, Cody Coleman, Bala Desinghu, Gregory Diamos, Debo Dutta, Diane Feddema, Grigori Fursin, Xinyuan Huang, Satyananda Kashyap, Nicholas Lane, Indranil Mallick, Pietro Mascagni, Virendra Mehta, Cassiano Ferro Moraes, Vivek Natarajan, Nikola Nikolov, Nicolas Padoy, Gennady Pekhimenko, Vijay Janapa Reddi, G. Anthony Reina, Pablo Ribalta, Abhishek Singh, Jayaraman J. Thiagarajan, Jacob Albrecht, Thomas Wolf, Geralyn Miller, Huazhu Fu, Prashant Shah, Daguang Xu, Poonam Yadav, David Talby, Mark M. Awad, Jeremy P. Howard, Michael Rosenthal, Luigi Marchionni, Massimo Loda, Jason M. Johnson, Spyridon Bakas, Peter Mattson
Summary: Medical artificial intelligence has great potential to enhance healthcare. MedPerf is an open platform that focuses on benchmarking AI models in the medical domain. It enables facilities to evaluate and verify the performance of AI models while prioritizing privacy.
NATURE MACHINE INTELLIGENCE
(2023)
Review
Medicine, General & Internal
Pranav Rajpurkar, Matthew P. Lungren
Summary: This article examines the advantages and limitations of current clinical radiologic AI systems, new clinical workflows, and the potential effect of generative AI and large multimodal foundation models.
NEW ENGLAND JOURNAL OF MEDICINE
(2023)
Article
Medicine, General & Internal
Jonathan J. Y. Heng, Desmond B. Teo, L. F. Tan
Summary: Artificial intelligence is rapidly advancing in the field of medicine. The emergence of Chat Generative Pre-trained Transformer (ChatGPT) has sparked both excitement and concerns among technology leaders, researchers, educators, and policy makers. AI is expected to revolutionize medicine with its potential benefits and limitations. Medical educators need to adapt and develop skills and curricula to effectively harness the innovative power of this new technology.
POSTGRADUATE MEDICAL JOURNAL
(2023)
Article
Computer Science, Artificial Intelligence
Andre M. Carrington, Douglas G. Manuel, Paul W. Fieguth, Tim Ramsay, Venet Osmani, Bernhard Wernly, Carol Bennett, Steven Hawken, Olivia Magwood, Yusuf Sheikh, Matthew McInnes, Andreas Holzinger
Summary: This paper proposes a new method called deep ROC analysis to evaluate the performance of binary classifiers and diagnostic tests. It provides more detailed information compared to traditional performance measures. The method measures the performance in multiple groups and allows comparisons between groups. The paper also offers a new interpretation of AUC as balanced average accuracy, relevant to individuals.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Biochemical Research Methods
Andreas Holzinger, Katharina Keiblingera, Petr Holub, Kurt Zatloukal, Heimo Mueller
Summary: Due to the successes of AI, such as ChatGPT, and the combination of biotechnology, new potential solutions have emerged to address global problems and contribute to sustainability.
Article
Radiology, Nuclear Medicine & Medical Imaging
Gabriel Adelsmayr, Michael Janisch, Heimo Mueller, Andreas Holzinger, Emina Talakic, Elmar Janek, Simon Streit, Michael Fuchsjaeger, Helmut Schoellnast
Summary: This study aimed to investigate whether CT texture analysis can differentiate between different types of lung cancers and tumors. The study included 133 patients who underwent CT-guided biopsy of the lung. The results showed significant differences in texture features between different entities, and the use of a HU threshold affected the results of the analysis. Rating: 7/10
EUROPEAN JOURNAL OF RADIOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Federico Cabitza, Andrea Campagner, Gianclaudio Malgieri, Chiara Natali, David Schneeberger, Karl Stoeger, Andreas Holzinger
Summary: This paper presents a framework for defining different types of explanations of AI systems and criteria for evaluating their quality. It proposes a structural view of constructing explanations and suggests a typology based on the explanandum, explanantia, and their relationship. The paper highlights the importance of epistemological and psychological perspectives in defining quality criteria and aims to support clear inventories, verification criteria, and validation methods for AI explainability.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Biophysics
Theresa Letz, Carina Hoerandtner, Matthias C. Braunisch, Peter Gundel, Julia Matschkal, Martin Bachler, Georg Lorenz, Andrea Koerner, Carolin Schaller, Moritz Lattermann, Andreas Holzinger, Uwe Heemann, Siegfried Wassertheurer, Christoph Schmaderer, Christopher C. Mayer
Summary: The aim of this study is to compare automatic to manual measurements of left ventricular hypertrophy (LVH) parameters and investigate their predictive value for cardiovascular and all-cause mortality in patients with end-stage kidney disease (ESKD). The study found that automatic algorithms can be as reliable as manual measurements in assessing LVH parameters and predicting risk in ESKD patients.
PHYSIOLOGICAL MEASUREMENT
(2023)
Review
Pathology
Markus Plass, Michaela Kargl, Tim-Rasmus Kiehl, Peter Regitnig, Christian Geissler, Theodore Evans, Norman Zerbe, Rita Carvalho, Andreas Holzinger, Heimo Mueller
Summary: The development of digital pathology allows pathologists to utilize AI-based computer programs for advanced analysis of whole slide images. However, the best-performing AI algorithms for image analysis are considered black boxes, making it unclear why they produce specific results. This article addresses the issue of explainability in digital pathology and discusses the necessity of explainable AI (XAI) techniques to enhance transparency and causal understanding. The authors argue for the development of user interfaces that enable contextual understanding and interactive questioning to bridge the gap between AI processes and medical experts' knowledge.
JOURNAL OF PATHOLOGY CLINICAL RESEARCH
(2023)
Article
Mathematics, Interdisciplinary Applications
Claire Jean-Quartier, Katharina Bein, Lukas Hejny, Edith Hofer, Andreas Holzinger, Fleur Jeanquartier
Summary: In response to socioeconomic development, this study focuses on the transparency and sustainability aspects of artificial intelligence in terms of energy consumption. The research measures carbon emissions and energy consumption of Python algorithms and tests the impact of explainability on algorithmic energy consumption. The results can guide the selection of tools to measure algorithmic energy consumption and raise awareness of emission-based model optimization by highlighting the sustainability of explainable artificial intelligence.
Article
Health Care Sciences & Services
Julian Matschinske, Julian Spaeth, Mohammad Bakhtiari, Niklas Probul, Mohammad Mahdi Kazemi Majdabadi, Reza Nasirigerdeh, Reihaneh Torkzadehmahani, Anne Hartebrodt, Balazs-Attila Orban, Sandor-Jozsef Fejer, Olga Zolotareva, Supratim Das, Linda Baumbach, Josch K. Pauling, Olivera Tomasevic, Bela Bihari, Marcus Bloice, Nina C. Donner, Walid Fdhila, Tobias Frisch, Anne-Christin Hauschild, Dominik Heider, Andreas Holzinger, Walter Hoetzendorfer, Jan Hospes, Tim Kacprowski, Markus Kastelitz, Markus List, Rudolf Mayer, Monika Moga, Heimo Mueller, Anastasia Pustozerova, Richard Roettger, Christina C. Saak, Anna Saranti, Herald H. H. W. Schmidt, Christof Tschohl, Nina K. Wenke, Jan Baumbach
Summary: Machine learning and artificial intelligence have achieved promising results in various fields, driven by the increasing availability of data. However, these data are often distributed across different institutions and cannot be easily shared due to strict privacy regulations. Federated learning (FL) enables the training of distributed machine learning models without sharing sensitive data. However, implementing FL is time-consuming and requires advanced programming skills and complex technical infrastructures.
JOURNAL OF MEDICAL INTERNET RESEARCH
(2023)
Review
Computer Science, Artificial Intelligence
Andreas Holzinger, Anna Saranti, Alessa Angerschmid, Bettina Finzel, Ute Schmid, Heimo Mueller
Summary: Artificial intelligence has made significant progress in standard pattern recognition tasks, but there is still a significant gap between AI and human-level concept learning. To analyze current approaches and drive progress, experimental environments and diagnostic/benchmark datasets are needed for explainable machine intelligence. This paper provides an overview of current AI solutions for benchmarking concept learning, reasoning, and generalization, discusses state-of-the-art diagnostic/benchmark datasets, and explores future research directions in this exciting field.
Article
Computer Science, Interdisciplinary Applications
Johann Steszgal, Peter Kieseberg, Andreas Holzinger
Summary: Reducing food waste is crucial for minimizing the human impact on the environment and improving the efficient use of natural resources. The healthcare sector, in particular, has significant potential for adopting sustainable practices in food management. This article highlights the key challenges and proposes a solution for reducing food waste in real-world healthcare settings.
Article
Radiology, Nuclear Medicine & Medical Imaging
Gabriel Adelsmayr, Michael Janisch, Ann-Katrin Kaufmann-Buehler, Magdalena Holter, Emina Talakic, Elmar Janek, Andreas Holzinger, Michael Fuchsjaeger, Helmut Schoellnast
Summary: The reproducibility problems in radiomics are well known, and the segmentation of target lesions greatly affects texture analysis variability. This study aimed to compare the interobserver reliability of manual 2D vs. 3D lung lesion segmentation with and without pre-definition of volume using a HU threshold.
EUROPEAN RADIOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Edgar R. Weippl, Andreas Holzinger, Peter Kieseberg
MACHINE LEARNING AND KNOWLEDGE EXTRACTION
(2023)
Article
Computer Science, Artificial Intelligence
Iztok Fister Jr, Iztok Fister, Vili Podgorelec, Sancho Salcedo-Sanz, Andreas Holzinger
Summary: This article introduces a novel visualization method developed based on the principles of explainable artificial intelligence to present association rules for time series. The experiments conducted in the context of smart agriculture demonstrate the great potential of the proposed visualization method.
