News Item
Multidisciplinary Sciences
Davide Castelvecchi
Summary: Computer scientists say that freely accessible large language models have accelerated the pace of innovation.
Article
Gastroenterology & Hepatology
Aaron I. F. Poon, Joseph J. Y. Sung
Summary: One of the biggest challenges in utilizing artificial intelligence (AI) in medicine is the lack of trust from both physicians and patients. Improving the interpretability of machine learning (ML) algorithms is crucial in successfully implementing AI in medicine. Opening the black box in AI medicine through a stepwise approach can help build trust and acceptance, ultimately advancing the development of AI technology in healthcare.
JOURNAL OF GASTROENTEROLOGY AND HEPATOLOGY
(2021)
Article
Oncology
Dalia Fahmy, Heba Kandil, Adel Khelifi, Maha Yaghi, Mohammed Ghazal, Ahmed Sharafeldeen, Ali Mahmoud, Ayman El-Baz
Summary: This manuscript discusses the applications of artificial intelligence (AI) in lung segmentation, pulmonary nodule segmentation, and classification. It highlights the importance of early detection of pulmonary nodules for lung cancer treatment and saving lives, and how AI technology can improve diagnostic accuracy.
Article
Computer Science, Hardware & Architecture
Lingjuan Lyu, Yitong Li, Karthik Nandakumar, Jiangshan Yu, Xingjun Ma
Summary: This article proposes a novel reputation system using digital tokens and local credibility to ensure fairness and privacy in collaborative deep learning. It also introduces a fair and differentially private decentralised deep learning framework, which uses a two-stage scheme to derive more accurate local models in a fair and private manner.
IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING
(2022)
Article
Computer Science, Artificial Intelligence
Waddah Saeed, Christian Omlin
Summary: Significant progress has been made in artificial intelligence (AI) in the past decade, but the increasing complexity and lack of transparency of black-box AI models remain challenges. To address this, Explainable AI (XAI) has been proposed to make AI more transparent and advance its adoption. This study provides a systematic meta-survey on the challenges and future research directions in XAI, organized into general challenges and research directions, as well as challenges and research directions based on the machine learning life cycle's phases: design, development, and deployment. It contributes to the XAI literature by offering a guide for future exploration.
KNOWLEDGE-BASED SYSTEMS
(2023)
Editorial Material
Multidisciplinary Sciences
Arthur Spirling
Summary: Researchers should steer clear of proprietary models and instead focus on creating transparent large language models for the sake of reproducibility.
Editorial Material
Computer Science, Artificial Intelligence
Frank Xing, Bjoern Schuller, Iti Chaturvedi, Erik Cambria, Amir Hussain
Summary: Neural network-based methods, such as word2vec and GPT-based models, have achieved significant progress in AI research, especially in handling large datasets. However, these methods lack in-depth understanding of the internal features and representations of the data, leading to various problems and concerns.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2023)
News Item
Multidisciplinary Sciences
Katharine Sanderson
Summary: Tools powered by large language models aim to assist researchers in comprehending and conducting scientific research.
Article
Multidisciplinary Sciences
Yin Li, Yueying Ni, Rupert A. C. Croft, Tiziana Di Matteo, Simeon Bird, Yu Feng
Summary: This study utilizes artificial intelligence to generate super-resolution versions of low-resolution cosmological Nbody simulations, enhancing the resolution and accurately replicating the high-resolution matter power spectrum and dark matter halo mass function in large-scale environments.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2021)
Review
Dermatology
Aditya K. Gupta, Iordanka A. Ivanova, Helen J. Renaud
Summary: Artificial intelligence applications in medicine, such as deep learning diagnostic models for skin lesions, are rapidly evolving, with new AI applications now emerging in hair restoration and disorder diagnosis. Current AI applications in hair restoration include fully automated systems for hair detection and growth measurement, while new deep learning-based systems are being proposed for scalp diagnosis and automated hair loss measurements, including devices for self-diagnosis. As these emerging technologies become more accessible to clinicians and patients, hair restoration experts should be aware of their potential benefits and limitations.
DERMATOLOGIC THERAPY
(2021)
Review
Urology & Nephrology
Yuankai Huo, Ruining Deng, Quan Liu, Agnes B. Fogo, Haichun Yang
Summary: The explosive growth of AI technologies, particularly in deep learning, has led to revolutionary applications in AI-assisted healthcare, including in renal pathology. However, successful integration of AI in renal pathology requires close interdisciplinary collaborations between computer scientists and renal pathologists. Understanding the high-level principles of AI technologies and optimizing AI techniques for renal pathology are crucial for future applications in this field.
KIDNEY INTERNATIONAL
(2021)
Article
Computer Science, Artificial Intelligence
Thomas P. Quinn, Stephan Jacobs, Manisha Senadeera, Vuong Le, Simon Coghlan
Summary: This article uses the analogy of the three Christmas ghosts to guide readers through the past, present, and future of medical AI. It highlights the reliance on opaque models in modern machine learning and discusses the implications for transparency in healthcare. The article argues that opaque models lack quality assurance, trust, and hinder physician-patient dialogue, and suggests upholding transparency in model design and validation to ensure the success of medical AI.
ARTIFICIAL INTELLIGENCE IN MEDICINE
(2022)
Article
Radiology, Nuclear Medicine & Medical Imaging
Almir Bitencourt, Isaac Daimiel Naranjo, Roberto Lo Gullo, Carolina Rossi Saccarelli, Katja Pinker
Summary: Advancements in imaging analysis and AI technology have opened up various potential applications in breast imaging, such as breast cancer risk prediction, lesion detection and classification, and radiogenomics. The use of AI tools in breast imaging presents an unprecedented opportunity to extract clinical value from imaging data.
EUROPEAN JOURNAL OF RADIOLOGY
(2021)
Review
Obstetrics & Gynecology
Tammy Lee, Jay Natalwala, Vincent Chapple, Yanhe Liu
Summary: With the advancement in computing power and accumulation of embryo image data, artificial intelligence (AI) has been applied in embryo selection in IVF. Machine learning (ML) has the potential to reduce subjectivity and save time, but the lack of interpretability in modern deep learning (DL) techniques has raised concerns. The effectiveness of black-box models lacks confirmation from randomized controlled trials, and recent evidence suggests underperformance compared to interpretable ML models. Interpretable AI is gaining support due to its ethical advantages and technical feasibility. A novel classification system for embryo selection models is proposed, focusing on subjectivity, interpretability, and explainability.
HUMAN REPRODUCTION
(2023)
Review
Ophthalmology
Ciara O'Byrne, Abdallah Abbas, Edward Korot, Pearse A. Keane
Summary: Automated deep learning allows users without coding expertise to develop deep learning algorithms, rapidly establishing itself as a valuable tool for those with limited technical experience. Despite residual challenges, it offers considerable potential in the future of patient management, clinical research and medical education.
CURRENT OPINION IN OPHTHALMOLOGY
(2021)