Less is More: Efficient Black-box Attribution via Minimal Interpretable Subset Selection

Open in new window