Katulu Federated Learning

The Game Changer for Industrial Machine Learning

Katulu Federated Learning (FL) creates entirely new opportunities for the use of Machine Learning in the industry. Decentralized machine learning across system boundaries creates more robust models for outstanding results with every customer - in line with data protection.

Take your digital value-added services to the next level and make machine learning combined with data protection your unique selling point.

Models learn from each other without sharing sensitive Data

Katulu FL trains the analysis models decentrally in the customer's production environment. The transfer of these models takes place exclusively in anonymized form - using the latest anonymization and encryption technologies. The exchange of sensitive raw data is completely eliminated - without sacrificing insights.

Customized Models for each Machine

With Katulu FL, a customized model can be trained for each machine and each customer's configuration (tooling, peripherals, material, etc.). This customized model is based on information from many different models. This allows Katulu to create robust models optimized for the customer's unique use case, even when the customer has only a limited amount of data.

Iterative Improvements

With each iteration, a central model is formed from the customers' analysis models. The combined findings of this model are continuously fed back into the individual customer models, thereby improving them. This iterative process increases the accuracy and robustness of the models and also avoids bias effects such as prejudice or distortion despite customer-specific optimization.