Timezone: »

Legendre Decomposition for Tensors
Mahito Sugiyama · Hiroyuki Nakahara · Koji Tsuda

Wed Dec 05 01:55 PM -- 02:00 PM (PST) @ Room 517 CD

We present a novel nonnegative tensor decomposition method, called Legendre decomposition, which factorizes an input tensor into a multiplicative combination of parameters. Thanks to the well-developed theory of information geometry, the reconstructed tensor is unique and always minimizes the KL divergence from an input tensor. We empirically show that Legendre decomposition can more accurately reconstruct tensors than other nonnegative tensor decomposition methods.

Author Information

Mahito Sugiyama (National Institute of Informatics)
Hiroyuki Nakahara (RIKEN Brain Science Institute)
Koji Tsuda (The University of Tokyo / RIKEN)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors