QUINT: Node embedding using network hashing

Representation learning using network embedding has received tremendous attention due to its efficacy solve downstream tasks. Popular methods (such as deepwalk, node2vec, LINE) are based on a neural architecture, thus unable scale large networks both in terms of time and space usage. Recently, we proposed BinSketch, sketching technique for compressing binary vectors vectors. In this paper, show how extend BinSketch use it hashing. Our proposal named QUINT is built upon embeds nodes sparse onto low-dimensional simple bi-wise operations. the first kind that provides gain speed usage without compromising much accuracy Extensive experiments conducted compare with seven state-of-the-art two end tasks - link prediction node classification. We observe huge performance speedup (up 7000x) saving 80x) bit-wise nature obtain embedding. Moreover, consistent top-performer among baselines across all datasets. empirical observations backed by rigorous theoretical analysis justify effectiveness QUINT. particular, prove retains enough structural information which can be used further approximate many topological properties high confidence.