English

Saddlepoints in Unsupervised Least Squares

Machine Learning 2021-04-13 v1 Machine Learning

Abstract

This paper sheds light on the risk landscape of unsupervised least squares in the context of deep auto-encoding neural nets. We formally establish an equivalence between unsupervised least squares and principal manifolds. This link provides insight into the risk landscape of auto--encoding under the mean squared error, in particular all non-trivial critical points are saddlepoints. Finding saddlepoints is in itself difficult, overcomplete auto-encoding poses the additional challenge that the saddlepoints are degenerate. Within this context we discuss regularization of auto-encoders, in particular bottleneck, denoising and contraction auto-encoding and propose a new optimization strategy that can be framed as particular form of contractive regularization.

Keywords

Cite

@article{arxiv.2104.05000,
  title  = {Saddlepoints in Unsupervised Least Squares},
  author = {Samuel Gerber},
  journal= {arXiv preprint arXiv:2104.05000},
  year   = {2021}
}
R2 v1 2026-06-24T01:03:09.187Z