Skip to main content

Intelligence Computing for Challenging Clinical Data: Robust, Practical and Trustworthy

Edited by Honghao Gao, Wenbing Zhao, and Yuyu Yin

Accuracy, efficiency and transferability are realistic issues in clinical practice, and research hotspots as well. Especially in developing countries, the number of clinicians is not enough to meet the population growth. One doctor has to diagnose dozens of patients every day, with traditional clinical methods. In recent years, artificial intelligence (AI) techniques, e.g., deep learning, tend to mature and have been applied in many clinical tasks successfully. Well-designed networks are able to detect various lesions in medical images, such as pulmonary nodule, diabetic retinopathy, and skin cancer. The accuracy and efficiency have also achieved or exceeded human experts. Introducing AI-aided methods to traditional clinical process would help reduce the burden of doctors, or improve diagnostic ability of inexperienced clinicians.

AI can be a sharp tool with sufficient data and deep architectures. But in many circumstances, models only give a simple result without any evidences, which is still unconvincing for clinical use. Moreover, challenging clinical data (small data, imbalanced data in rare diseases, unobvious lesions in early stage, multiple data source with different distributions, etc.) that cannot meet the requirements of AI algorithms are common. Existing intelligent solutions may be unstable or fail for those data. This requires novel computing techniques with robust, practical and trustworthy.

This Special Issue aims to invite original research papers that tackle challenging clinical data and scenarios, including theoretical research, practical models improvement, clinical-oriented data analysis, deployed digital system, and farsighted intelligent architecture. Submitted papers should clarify the substantively different from work that has already been published, or accepted for publication, or submitted in parallel to other conferences or journals. 

There are currently no articles in this collection.

Annual Journal Metrics