Artificial intelligence (A.I.) is expected to significantly influence the practice of medicine and the delivery of healthcare in the near future. While there are only a handful of practical examples for its medical use with enough evidence, hype and attention around the topic are significant. There are so many papers, conference talks, misleading news headlines and study interpretations that a short and visual guide any medical professional can refer back to in their professional life might be useful. For this, it is critical that physicians understand the basics of the technology so they can see beyond the hype, evaluate A.I.-based studies and clinical validation; as well as acknowledge the limitations and opportunities of A.I. This paper aims to serve as a short, visual and digestible repository of information and details every physician might need to know in the age of A.I. We describe the simple definition of A.I., its levels, its methods, the differences between the methods with medical examples, the potential benefits, dangers, challenges of A.I., as well as attempt to provide a futuristic vision about using it in an everyday medical practice.

Our pipeline consists of a hand-crafted preprocessor and a neural network classifier. We applied transformations on the physiologic signals to gain features in both time-and frequency domains. The proposed algorithm was trained on 994 annotated records of polysomnographic signals. Most of the features were generated from the EEG signal such as power spectral density, and entropy. We extracted features from the EOG, EMG, airflow, and ECG signals too. All the features were normalized. These 68 features were resampled in 21 non-continuous moments around the current timestamp, and fed into a 3layer neural network in order to assign a probability of arousal at each second. Arousal samples were enriched during training to battle data imbalance. Additional (auxiliary) losses can guide the network to learn high-level concepts, even though they will not be evaluated. We used sleep stages as additional training targets, which were easier to learn than arousals despite being multi-class. This approach slightly increased arousal AUPRC. Our submitted results for the entire test set were evaluated: AUPRC=0.42. Our 10-fold cross validation results for the AUPRC are the following: [

No abstract

BACKGROUND Background: Artificial intelligence (AI)/machine learning (ML)-based medical devices and algorithms are rapidly changing the medical field. To provide an insight into the trends in AI and ML in healthcare, we conducted an international patent analysis. OBJECTIVE - METHODS Methods: A systematic patent analysis, focusing on AI/ML-based patents in healthcare, was performed using the Espacenet database (from 01-2012 until 07-2022). This database includes patents from the China National Intellectual Property Administration (CNIPA), European Patent Office (EPO), Japan Patent Office (JPO), Korean Intellectual Property Office (KIPO), United States Patent and Trademark Office (USPTO). RESULTS Results: We identified 10967 patents: 7332 (66.9%) from CNIPA, 191 (1.7%) from EPO, 163 (1.5%) from JPO, 513 (4.7%) from KIPO and 2768 (25.2%) from USPTO. The number of published patents showed a yearly doubling from 2015 until 2021. Five international companies had the greatest impact on this increase: Ping An Medical and Healthcare Management with 568 (5.2%) patents, Siemens Healthineers with 273 (2.5%), IBM with 226 (2.1%), Philips Healthcare with 150 (1.4%) and Shanghai United Imaging Healthcare with 144 (1.3%). CONCLUSIONS Conclusion: This international patent analysis showed a linear increase in patents published by the five largest patent offices. An open access database with interactive search options was launched for AI/ML-based patents in healthcare.

Background Artificial intelligence (AI)– and machine learning (ML)–based medical devices and algorithms are rapidly changing the medical field. To provide an insight into the trends in AI and ML in health care, we conducted an international patent analysis. Objective It is pivotal to obtain a clear overview on upcoming AI and MLtrends in health care to provide regulators with a better position to foresee what technologies they will have to create regulations for, which are not yet available on the market. Therefore, in this study, we provide insights and forecasts into the trends in AI and ML in health care by conducting an international patent analysis. Methods A systematic patent analysis, focusing on AI- and ML-based patents in health care, was performed using the Espacenet database (from January 2012 until July 2022). This database includes patents from the China National Intellectual Property Administration, European Patent Office, Japan Patent Office, Korean Intellectual Property Office, and the United States Patent and Trademark Office. Results We identified 10,967 patents: 7332 (66.9%) from the China National Intellectual Property Administration, 191 (1.7%) from the European Patent Office, 163 (1.5%) from the Japan Patent Office, 513 (4.7%) from the Korean Intellectual Property Office, and 2768 (25.2%) from the United States Patent and Trademark Office. The number of published patents showed a yearly doubling from 2015 until 2021. Five international companies that had the greatest impact on this increase were Ping An Medical and Healthcare Management Co Ltd with 568 (5.2%) patents, Siemens Healthineers with 273 (2.5%) patents, IBM Corp with 226 (2.1%) patents, Philips Healthcare with 150 (1.4%) patents, and Shanghai United Imaging Healthcare Co Ltd with 144 (1.3%) patents. Conclusions This international patent analysis showed a linear increase in patents published by the 5 largest patent offices. An open access database with interactive search options was launched for AI- and ML-based patents in health care.