Timezone: »

Non-linear Metric Learning
Dor Kedem · Stephen Tyree · Kilian Q Weinberger · Fei Sha · Gert Lanckriet

Thu Dec 06 02:00 PM -- 12:00 AM (PST) @ Harrah’s Special Events Center 2nd Floor

In this paper, we introduce two novel metric learning algorithms, χ2-LMNN and GB-LMNN, which are explicitly designed to be non-linear and easy-to-use. The two approaches achieve this goal in fundamentally different ways: χ2-LMNN inherits the computational benefits of a linear mapping from linear metric learning, but uses a non-linear χ2-distance to explicitly capture similarities within histogram data sets; GB-LMNN applies gradient-boosting to learn non-linear mappings directly in function space and takes advantage of this approach's robustness, speed, parallelizability and insensitivity towards the single additional hyper-parameter. On various benchmark data sets, we demonstrate these methods not only match the current state-of-the-art in terms of kNN classification error, but in the case of χ2-LMNN, obtain best results in 19 out of 20 learning settings.

Author Information

Dor Kedem (Washington University in St. Louis)
Stephen Tyree (Washington University in St. Louis)
Kilian Q Weinberger (Cornell University / ASAPP Research)
Fei Sha (University of Southern California (USC))
Gert Lanckriet (U.C. San Diego)

More from the Same Authors