We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Tensor Train Neighborhood Preserving Embedding

Abstract: In this paper, we propose a Tensor Train Neighborhood Preserving Embedding (TTNPE) to embed multi-dimensional tensor data into low dimensional tensor subspace. Novel approaches to solve the optimization problem in TTNPE are proposed. For this embedding, we evaluate novel trade-off gain among classification, computation, and dimensionality reduction (storage) for supervised learning. It is shown that compared to the state-of-the-arts tensor embedding methods, TTNPE achieves superior trade-off in classification, computation, and dimensionality reduction in MNIST handwritten digits and Weizmann face datasets.
Comments: Accepted to IEEE Transactions on Signal Processing, Mar 2018
Subjects: Machine Learning (cs.LG); Information Theory (cs.IT); Machine Learning (stat.ML)
DOI: 10.1109/TSP.2018.2816568
Cite as: arXiv:1712.00828 [cs.LG]
  (or arXiv:1712.00828v2 [cs.LG] for this version)

Submission history

From: Vaneet Aggarwal [view email]
[v1] Sun, 3 Dec 2017 20:09:48 GMT (306kb)
[v2] Sat, 10 Mar 2018 01:11:03 GMT (1716kb,D)

Link back to: arXiv, form interface, contact.