Regularization of Mixture Models for Robust Principal Graph Learning

@article{Bonnaire2021RegularizationOM,
  title={Regularization of Mixture Models for Robust Principal Graph Learning},
  author={Tony Bonnaire and Aur{\'e}lien Decelle and Nabila Aghanim},
  journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year={2021},
  volume={PP},
  pages={1-1}
}
A regularized version of Mixture Models is proposed to learn a principal graph from a distribution of D-dimensional data points. In the particular case of manifold learning for ridge detection, we assume that the underlying structure can be modeled as a graph acting like a topological prior for the Gaussian clusters turning the problem into a maximum a posteriori estimation. Parameters of the model are iteratively estimated through an Expectation-Maximization procedure making the learning of… 

Figures from this paper

References

SHOWING 1-10 OF 39 REFERENCES

Regularization-free principal curve estimation

TLDR
This paper introduces a new objective function, facilitated through a modification of the principal curve estimation approach, for which all critical points are principal curves and minima, and removes the fundamental issue for model selection in principal curves estimation.

Nonparametric Ridge Estimation

TLDR
Ridge estimation is an extension of mode finding and is useful for understanding the structure of a density and can be used to find hidden structure in point cloud data.

Robust and Scalable Learning of Complex Intrinsic Dataset Geometry via ElPiGraph

TLDR
ElPiGraph exploits and further develops the concept of elastic energy, the topological graph grammar approach, and a gradient descent-like optimization of the graph topology, and is capable of approximating data point clouds via principal graph ensembles.

Laplacian Eigenmaps for Dimensionality Reduction and Data Representation

TLDR
This work proposes a geometrically motivated algorithm for representing the high-dimensional data that provides a computationally efficient approach to nonlinear dimensionality reduction that has locality-preserving properties and a natural connection to clustering.

Locally Defined Principal Curves and Surfaces

TLDR
A novel theoretical understanding of principal curves and surfaces, practical algorithms as general purpose machine learning tools, and applications of these algorithms to several practical problems are presented.

Generalized Mode and Ridge Estimation

TLDR
A method is proposed for studying the geometric structure of generalized densities using a modification of the mean shift algorithm and its variant, subspace constrained mean shift, which can be used to perform clustering and to calculate a measure of connectivity between clusters.

Laplacian Eigenmaps and Spectral Techniques for Embedding and Clustering

TLDR
The algorithm provides a computationally efficient approach to nonlinear dimensionality reduction that has locality preserving properties and a natural connection to clustering.

Geometric Inference for Probability Measures

TLDR
Replacing compact subsets by measures, a notion of distance function to a probability distribution in ℝd is introduced and it is shown that it is possible to reconstruct offsets of sampled shapes with topological guarantees even in the presence of outliers.

Principal Graphs and Manifolds

TLDR
This chapter gives a brief practical introduction into the methods of construction of general principal objects, i.e. objects embedded in the ‘middle’ of the multidimensional data set, using the family of expectation/maximisation algorithms with nearest generalisations.

Nonlinear dimensionality reduction by locally linear embedding.

TLDR
Locally linear embedding (LLE) is introduced, an unsupervised learning algorithm that computes low-dimensional, neighborhood-preserving embeddings of high-dimensional inputs that learns the global structure of nonlinear manifolds.