The quality of algorithms is predominantly related to the amount of data available. In recent years, this availability has increased rapidly. In particular, the Internet of Things (IoT) generates large amounts of data from millions of devices. This makes it possible to train increasingly advanced Machine Learning models. In practice, however, making this data available to train a centralized model can be problematic due to regulatory restrictions or technical hurdles in transmitting large amounts of data over low bandwidths. One solution to these challenges is Federated Learning.
Federated Learning – An Answer for Companies in Terms of Artificial Intelligence (AI)
Here, all training data is stored exclusively on local devices or clients and model training is decentralized. Clients receive a model and improve it by learning from their local data. A summary of a client’s updated model is sent to a server via encrypted communication. There, it is aggregated with updates from other clients to form a common global model, that is, fused into a single global model. This model is then sent to the clients again, and the training process continues until the model has been sufficiently trained.
Successful Use of Federated Learning
The best-known example of federated learning is text support in smartphones. There are also countless other applications, such as an app for predicting a stroke or supporting the prognosis of cancer. There are also many examples in the area of billing auditing of decentralized billing data, e.g. in the healthcare sector, or marketing or individual risk analysis.
Federated Learning Project
We have already successfully used Federated Learning as part of the »Bauhaus Mobility Lab« (BML) project. The goal of the project was to build a digital lab platform that combines AI technologies with data. The project is being funded by the German Federal Ministry of Economics and Climate Protection (BMWK) until 2023 as part of the innovation competition »Artificial Intelligence as a Driver of Economically Relevant Ecosystems«.
We are currently looking for companies and partners with whom we can realise our workshops individually adapted to the use cases and previous knowledge.
Sounds interesting? More information about our workshops can be found here.