No abstract

Despite substantial gains facilitated by Artificial Intelligence (AI) in recent years, it has to be applied very cautiously in sensitive domains like medicine due to the lack of explainability of many methods in this field. We aim to provide a system to overcome these issues of medical AI applications by means of our concept of medical operational AI detailed in this paper. We make use of various methods of AI and utilize knowledge graphs in particular. The latter is continuously updated by medical experts based on medical literature such as peer-reviewed papers and standard online sources such as UpToDate. We thoroughly derive a multi-level system tackling the corresponding challenges. In particular, its design encompasses (i) holistic diagnostic assistance on a macro level, (ii) predicitions and detailed suggestions for specific medical domains on a micro level, as well as (iii) AI-based optimizations of the overall system on a meta level. We detail practical merits of medical operational AI and discuss the state of the art beyond our solution.

Medical diagnosis is the process of making a prediction of the disease a patient is likely to have, given a set of symptoms and observations. This requires extensive expert knowledge, in particular when covering a large variety of diseases. Such knowledge can be coded in a knowledge graph -encompassing diseases, symptoms, and diagnosis paths. Since both the knowledge itself and its encoding can be incomplete, refining the knowledge graph with additional information helps physicians making better predictions. At the same time, for deployment in a hospital, the diagnosis must be explainable and transparent. In this paper, we present an approach using diagnosis paths in a medical knowledge graph. We show that those graphs can be refined using latent representations with RDF2vec, while the final diagnosis is still made in an explainable way. Using both an intrinsic as well as an expert-based evaluation, we show that the embedding-based prediction approach is beneficial for refining the graph with additional valid conditions.