For billing of medical services hospitals need to code the patient records with international codes. This task occupies many highly skilled professionals and binds a lot of resources. Furthermore, every hospital uses its own vocabulary for the description of diagnoses and treatments. The TLDia project examines the question of how the coding can be automated. The coding system should support the coding of handwritten doctor’s letters. Additionally, by using of transfer learning methods it should allow a transmission of the whole coding system between the hospitals or insurances. Transfer learning moves more and more into the focus of artificial intelligence. AI-systems need to be trained for every single task and require a high amount of training data. By application of transfer learning methods pre-trained systems can be used for execution of new tasks and the learned procedures can be transferred into another context.
Funding and Partners
The project TLDia is a cooperation with THM Friedberg (consortium leadership, cmp. more detailed project description on THM's Web page in German language) and the company Minds-Medical. It has received a two-year funding from the Hessian„Landes-Offensive zur Entwicklung Wissenschaftlich-ökonomischer Exzellenz“ (LOEWE) under grant no. 701/19-21.