Signed Graph Metric Learning via Gershgorin Disc Perfect Alignment.

@article{Yang2021SignedGM,
  title={Signed Graph Metric Learning via Gershgorin Disc Perfect Alignment.},
  author={Cheng Yang and Gene Cheung and Wei Hu},
  journal={IEEE transactions on pattern analysis and machine intelligence},
  year={2021},
  volume={PP}
}
Given a convex and differentiable objective [Formula: see text] for a real symmetric matrix [Formula: see text] in the positive definite (PD) cone, we propose a fast general metric learning framework that is entirely projection-free. We first assume that [Formula: see text] resides in a space [Formula: see text] of generalized graph Laplacian matrices corresponding to balanced signed graphs. Unlike low-rank metric matrices common in the literature, [Formula: see text] includes the important… 
Projection-free Graph-based Classifier Learning using Gershgorin Disc Perfect Alignment
TLDR
This paper proposes a fast projection-free method by solving a sequence of linear programs (LP) by recast the SDP relaxation to its SDP dual, where a feasible solution H 0 can be interpreted as a Laplacian matrix corresponding to a balanced signed graph sans the last node.
Fast Graph-based Binary Classifier Learning via Further Relaxation of Semi-Definite Relaxation
TLDR
This paper proposes a fast graph-based binary classification problem solver with only time complexity, which first derive an upper bound of the Shur complement of the last diagonal entry of the PSD classification matrix, and then derive a trivial solution of the dual of the semi-definite program (SDP) for the graph- Based Binary classification problem.
Unfolding Projection-free SDP Relaxation of Binary Graph Classifier via GDPA Linearization
TLDR
This paper unrolls a projection-free algorithm for semi-definite programming relaxation (SDR) of a binary graph classifier, where the PSD cone constraint is replaced by a set of “tightest possible” linear constraints per iteration.
Unsupervised Graph Spectral Feature Denoising for Crop Yield Prediction
TLDR
Experimental results from collected USDA data show that using denoised features as input, performance of a crop yield prediction model can be improved noticeably.
Manifold Graph Signal Restoration using Gradient Graph Laplacian Regularizer
TLDR
This paper generalizes GLR to gradient graph Laplacian regularizer (GGLR) that provably promotes piecewise planar (PWP) signal reconstruction and derives the means-square-error minimizing weight parameter for GGLR efficiently, trading off bias and variance of the signal estimate.
Fast Computation of Generalized Eigenvectors for Manifold Graph Embedding
TLDR
K-dimensional latent vectors for the N graph nodes are the first K generalized eigenvectors for (A, B), computed in $O(N)$ using LOBPCG, where K ≪ N.
Fast & Robust Image Interpolation Using Gradient Graph Laplacian Regularizer
TLDR
This work generalizes GLR to gradient graph Laplacian regularizer (GGLR) that provably promotes piecewise planar (PWP) signal reconstruction for the image interpolation problem—a 2D grid with random missing pixels that requires completion.

References

SHOWING 1-10 OF 48 REFERENCES
Graph Metric Learning via Gershgorin Disc Alignment
  • Cheng Yang, Gene Cheung, Wei Hu
  • Computer Science
    ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
  • 2020
TLDR
It is proved that left-ends of the Gershgorin discs can be aligned perfectly using the first eigenvector v of M, which is updated iteratively using Locally Optimal Block Preconditioned Conjugate Gradient (LOBPCG) with warm start as diagonal / off-diagonal terms are optimized.
Feature Graph Learning for 3D Point Cloud Denoising
TLDR
This work alternately optimize the diagonal and off-diagonal entries of a Mahalanobis distance matrix and constrain the Schur complement of sub-matrix to be positive definite (PD) via linear inequalities derived from the Gershgorin circle theorem.
Fast Graph Sampling Set Selection Using Gershgorin Disc Alignment
TLDR
The dual problem is reinterpreted as an intuitive disc coverage problem bearing strong resemblance to the famous NP-hard set cover (SC) problem and a fast approximation scheme is derived from a known SC error-bounded approximation algorithm.
Introduction to Graph Signal Processing
TLDR
Spectral analysis of graphs is discussed next and some simple forms of processing signal on graphs, like filtering in the vertex and spectral domain, subsampling and interpolation, are given.
Reconstruction-cognizant Graph Sampling Using Gershgorin Disc Alignment
TLDR
Experiments on real graph data show that the proposed algorithm can effectively promote large eigenvalue lower bounds, and the reconstruction MSE is the same or smaller than existing sampling methods for different budget K at much lower complexity.
Alternating Binary Classifier and Graph Learning from Partial Labels
TLDR
Improve the performance of such a graph-based classifier by simultaneously optimizing the feature weights used in the construction of the similarity graph by first formulating a boolean quadratic program with a graph signal smoothness objective, then relaxing it to a convex semi-definite program, solvable in polynomial time.
Graph Signal Processing: Overview, Challenges, and Applications
TLDR
An overview of core ideas in GSP and their connection to conventional digital signal processing are provided, along with a brief historical perspective to highlight how concepts recently developed build on top of prior research in other areas.
Learning Local Metrics and Influential Regions for Classification
TLDR
A new intuitive distance is defined with local metrics and influential regions, and subsequently a novel local metric learning algorithm called LMLIR is proposed for distance-based classification.
Distance Metric Learning for Large Margin Nearest Neighbor Classification
TLDR
This paper shows how to learn a Mahalanobis distance metric for kNN classification from labeled examples in a globally integrated manner and finds that metrics trained in this way lead to significant improvements in kNN Classification.
...
...