DynamicSparseNetworkforTimeSeries Classification: Learning Whatto" See "

In this paper, we propose a dynamic sparse network (DSN) with sparse connections for TSC, which can learn to cover various RF without cumbersome hyper-parameters tuning.