Hyperbolic Personalized Tag Recommendation

Personalized Tag Recommendation (PTR) aims to automat-
ically generate a list of tags for users to annotate web resources, the
so-called items, according to users’ tagging preferences. The main chal-
lenge of PTR is to learn representations of involved entities (i.e., users,
items, and tags) from interaction data without loss of structural proper-
ties in original data. To this end, various PTR models have been devel-
oped to conduct representation learning by embedding historical tag-
ging information into low-dimensional Euclidean space. Although such
methods are effective to some extent, their ability to model hierarchy,
which lies in the core of tagging information structures, is restricted by
Euclidean space’s polynomial expansion property. Since hyperbolic space
has recently shown its competitive capability to learn hierarchical data
with lower distortion than Euclidean space, we propose a novel PTR
model that operates on hyperbolic space, namely HPTR. HPTR learns
the representations of entities by modeling their interactive relationships
in hyperbolic space and utilizes hyperbolic distance to measure semantic
relevance between entities. Specially, we adopt tangent space optimiza-
tion to update model parameters. Extensive experiments on real-world
datasets have shown the superiority of HPTR over state-of-the-art base-
lines.