Article Series: Federated Machine Learning and its Application in the Finance Industry

The ADGM Academy Research Centre sat down with Professor Jin-Chuan Duan, Head of the Asian Institute of Digital Finance, at the National University of Singapore, and Richard Hills, Associate Managing Director at K2 Integrity, based in Abu Dhabi, to explore potential applications of federated learning in the financial industry, while taking a broader view on the evolution of technology and the barriers to adopting innovative solutions.

Without explicitly exchanging data samples, federated learning tries to train a machine learning algorithm using a variety of regional datasets found in local nodes. The fundamental idea is to train models on small samples of local data and distribute parameters (like the weights and biases of a deep neural network) among several local nodes. At minimum, the benefits that federated machine learning models could bring to the industry would enable a level playing field and provide equal benefit for all participants.