TIVITY-AI

TIVITY-AI

AI-based system for enabling demand-driven digitization processes in SMEs

TIVITY-AI

TIVITY-AI is an extension of the low code platform TITITY, which allows lay users to integrate Machine Learning into their application. Fortiss is involved into creating a user-centric workflow, and enabling lay users to create AI applications from a technical and interface point of view.

Project description

TIVITY is a low-code platform, which enables lay users to build custom applications without the need for professional development support. Low code platforms intend to close the gap between the necessity of digitalization and process automation, especially in smaller companies, and the lack of professionals in the IT sector.

TIVITY AI is meant to extent the possibilities of low code platforms by machine learning features. In low code platforms, the target domain and use case is not predetermined, but decision of the low code programmer, who often has limited technical and development knowledge. Therefore, special effort needs to be put into enabling the low code programmer to build an application which meets the end user’s needs. This means the low code programmer needs to be able to examine what’s possible with AI, which training and test data to use and how to debug the model output. Additionally, support for appropriate user interface design within human-AI interaction needs to be provided.

Research contribution

On the aspect of human-computer-interaction, fortiss will provide guidance for machine teaching in low code platforms. As an underlying, fortiss will therefore define mental models of lay users for document classification and content extraction within the machine teaching process.

Based on this, concepts for user guidance within the low-code development process will be created and evaluated. Here a special focus will be put on providing useful training and test data, as well as enabling debuggability for lay users.

Project duration

15.08.2021 – 14.08.2023

Project partner