DBIS

This Master thesis examines the impact of using Generative Replay-based methods on reducing Catastrophic Forgetting in Distributed Analytics applied to healthcare data. The study will focus on investigating how these methods can help preserve the knowledge gained from previously seen data distributions and prevent the model from forgetting this knowledge when learning on new data distributions in a distributed environment. The results of this research will contribute to the understanding of how Generative Replay can improve the performance of machine learning models in healthcare data analysis and could have implications for the development of more effective and reliable machine learning systems for this domain.