Ricerc@Sapienza

Recommender systems are an integral part of many web applica-
tions. With increasingly larger user bases, scalability has become an
important issue. Many of the most scalable algorithms with respect
to both space and running times are based on locality-sensitive
hashing (LSH). However, a significant drawback is that these meth-
ods are only able to handle insertions to user profiles and tend to
perform poorly when items may be removed. We initiate the study
of scalable locality-sensitive hashing for dynamic input. Specifi-