The XAI toolbox developed by Fraunhofer IOSB is designed with AI explainability at its core. It enables the swift evaluation of various XAI methods for AI procedures. That is, the toolbox can be used, for example, for data analysis, debugging, and explaining the prediction of any black-box model. This ensures trustworthiness in AI decisions. Currently, it supports scikit and TensorFlow based time series classifiers.
The XAI toolbox opens up the AI methods (black box) and shows the effects of each input on the outputs. For this purpose, it generates two kinds of explanations: local explanation for a single instance and global explanation for multiple instances or the entire model. On the one hand, this improves the developers ability to better explain the results, and on the other hand, it strengthens the users confidence in the methods. In addition, the toolbox has a user-friendly interface to support the use of AI models. This allows the user to intuitively and correctly interpret the result of the AI method and its explanations, and quickly draw the right conclusions. In addition, the user can evaluate different XAI methods and choose the most appropriate method for their application. Therefore, the XAI toolbox is a necessary platform for use in decision-critical domains such as medical, automotive, maritime, etc.