Medical Knowledge-Assisted Machine Learning Technologies in Individualized Medicine Volume III

This Research Topic is Volume III of a series. The previous volumes can be found here: Volume I and Volume II.

Machine learning has recently made impressive advances in applications ranging from computer vision to natural language processing and has been extensively used in bioinformatics and medical image analysis. However, a machine learning paradigm solely relying on data-driven connectionism may not always adequately address complex medical problems. On the other hand, medical experience based on symbolic representation and expert reasoning is rich in knowledge. The effective representation and embedding of medical knowledge can assist the optimization of machine learning algorithms, and effectively improve its performance, interpretability, and reliability. The advent of large-scale model technology, exemplified by ChatGPT, has underscored its advantages in addressing specific challenges within the medical domain, essentially embodying a form of knowledge extraction and application. Recognizing these advances therefore, the integration of disease knowledge with machine learning, particularly when enhanced by the sophisticated capabilities of large-scale models, presents a novel scientific research paradigm that could significantly enhance the effectiveness of machine learning applications in personalized medicine through enriched deep knowledge representation and expert-level reasoning.

Through this Research Topic, we aim to expand our understanding of knowledge-embedded machine learning in solving challenges in individualized medicine. We encourage submissions that demonstrate advancements using both traditional and large-scale model approaches, in genomic approaches and medical image analysis. Specifically, we are looking for:

Development and application of machine learning methods in cancer genomics, which may include utilizing large-scale models or traditional strategies to refine cancer risk evaluation.

Enhancement of image analysis in medical fields through machine learning, where submissions can demonstrate the use of either established techniques or explore novel applications with large-scale models for diagnosis and risk evaluation.

Innovative work in multi-omics analysis or integrated genomic and image data approaches, highlighting methods that either adhere to established practices or incorporate large-scale machine learning insights for improved integration or interpretability.

Novel contributions to translational medicine tools and software, with an openness to various methodologies including but not limited to the use of large-scale machine learning for processing high-throughput sequencing data and analyzing radiological and histological images.