Abstract: Learning hash functions for approximate nearest neighbor search of high-dimensional data has received a surge of interests in recent years. Most existing methods are often concerned with ...