a linear operator corresponding to a single layer, we develop a simple prun-ing method, coined lookahead pruning, by extending the single layer optimizationto a multi-layer o... particularly in the high-sparsity regime. Seehttps://github.com/alinlab/lookahead_pruning for codes.1 I NTRODUCTIONThe magnitude-equals-saliency approach has been lon...
