Deep learning model that integrates spectral embedding to partition a graph such that the normalized cut is the smallest possible. Based on Deep Learning and Spectral Embedding for Graph Partitioning.