Deep Learning in the laboratory environment is still limited to particular cases. For example, it is used to evaluate NIR spectra, detect tumors in tissues, or malfunction in technical equipment through sensors. All previous approaches have in common that a generalization is missing in the sense of a universal approach. Consequently, the user can't configure his solutions by simple hand movements.
ANNA starts precisely here, as the software learns from already categorized (previously interpreted by humans) series of measurements, analyzes unknown ones afterwards and creates laboratory reports based on configured record catalogs, comparable in their quality to those of experienced employees. ANNA thus closes the gap between the pure recognition of patterns in measurement data and the necessary generation of comprehensible laboratory reports for users and customers.