Automatic online multi-source domain adaptation

Knowledge transfer across several streaming processes remain challenging problem not only because of different distributions each stream but also rapidly changing and never-ending environments data streams. Albeit growing research achievements in this area, most existing works are developed for a single source domain which limits its resilience to exploit multi-source domains being beneficial recover from concept drifts quickly avoid the negative problem. An online adaptation technique under processes, namely automatic (AOMSDA), is proposed paper. The strategy AOMSDA formulated coupled generative discriminative approach denoising autoencoder (DAE) where central moment discrepancy (CMD)-based regularizer integrated handle existence thereby taking advantage complementary information sources. asynchronous place at time periods addressed by self-organizing structure node re-weighting strategy. Our numerical study demonstrates that capable outperforming counterparts 5 8 cases while ablation depicts learning component. In addition, general any number code shared publicly https://github.com/Renchunzi-Xie/AOMSDA.git.