There is a strong trend to reduce the complexity of design. This often works perfectly fine in most consumer apps. For the development of more complex research / software engineering projects, the number of functions cannot be reduced and complex use cases cannot be omitted.
It is clear that existing complexity should be simplified as much as possible.
The Cognitive Workbench is a new approach to analysis, knowledge extraction and training of recognizers based on AI algorithms. This an unknown system for many and often very complex. Since a revision of the code was planned, we used the opportunity to bring user-acquired knowledge into the product. We reworked the previous CWB concept in terms of both interface and improved workflows and interaction paradigms.
The result of the project was a completely revised technical code version of the Cognitive Workbench.
A redesigned appearance and improved workflows simplify the user's tasks and goals: build and configure cognitive computing applications to use state-of-the-art Natural Language Processing and Machine Learning algorithms. A standardized style guide and a component library as a side project that simplify and shorten the design and the frontend development process in the future.