Timezone: »
Convex Optimization for Blind Source Separation on a Statistical Manifold
Simon Luo · lamiae azizi · Mahito Sugiyama
We present a novel blind source separation (BSS) method using a hierarchical structure of sample space that is incorporated with a log-linear model. Our approach is formulated as a convex optimization with theoretical guarantees to uniquely recover a set of source signals by minimizing the KL divergence from a set of mixed signals. Source signals, received signals, and mixing matrices are realized as different layers in our hierarchical sample space. Our empirical results have demonstrated superiority compared to well established techniques.
Author Information
Simon Luo (The University of Sydney)
lamiae azizi (School of Mathematics and Statistics, The university of sydney)
Mahito Sugiyama (National Institute of Informatics)
More from the Same Authors
-
2021 : Detecting Poisoning Nodes in Federated Learning by Ranking Gradients »
Wanchuang Zhu · Benjamin Zhao · Simon Luo · Ke Deng -
2021 Poster: Fast Tucker Rank Reduction for Non-Negative Tensors Using Mean-Field Approximation »
Kazu Ghalamkari · Mahito Sugiyama -
2020 : Contributed Talk 1: The Volume of Non-Restricted Boltzmann Machines and Their Double Descent Model Complexity »
Prasad Cheema · Mahito Sugiyama -
2020 : Towards Geometric Understanding of Low-Rank Approximation »
Mahito Sugiyama · Kazu Ghalamkari -
2018 Poster: Legendre Decomposition for Tensors »
Mahito Sugiyama · Hiroyuki Nakahara · Koji Tsuda -
2018 Spotlight: Legendre Decomposition for Tensors »
Mahito Sugiyama · Hiroyuki Nakahara · Koji Tsuda -
2015 Poster: Halting in Random Walk Kernels »
Mahito Sugiyama · Karsten Borgwardt -
2013 Poster: Rapid Distance-Based Outlier Detection via Sampling »
Mahito Sugiyama · Karsten Borgwardt