ML Anthology

Robust Boltzmann Machines for Recognition and Denoising

Yichuan Tang, Ruslan Salakhutdinov, Geoffrey E. Hinton
CVPR 2012 pp. 2264-2271
doi:10.1109/CVPR.2012.6247936 /cvpr/2012/tang2012cvpr-robust/

Abstract

While Boltzmann Machines have been successful at unsupervised learning and density modeling of images and speech data, they can be very sensitive to noise in the data. In this paper, we introduce a novel model, the Robust Boltzmann Machine (RoBM), which allows Boltzmann Machines to be robust to corruptions. In the domain of visual recognition, the RoBM is able to accurately deal with occlusions and noise by using multiplicative gating to induce a scale mixture of Gaussians over pixels. Image denoising and in-painting correspond to posterior inference in the RoBM. Our model is trained in an unsupervised fashion with unlabeled noisy data and can learn the spatial structure of the occluders. Compared to standard algorithms, the RoBM is significantly better at recognition and denoising on several face databases.

PDF CVPR Semantic Scholar

Cite

Text

Tang et al. "Robust Boltzmann Machines for Recognition and Denoising." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2012. doi:10.1109/CVPR.2012.6247936

Markdown

[Tang et al. "Robust Boltzmann Machines for Recognition and Denoising." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2012.](https://mlanthology.org/cvpr/2012/tang2012cvpr-robust/) doi:10.1109/CVPR.2012.6247936

BibTeX

@inproceedings{tang2012cvpr-robust,
  title     = {{Robust Boltzmann Machines for Recognition and Denoising}},
  author    = {Tang, Yichuan and Salakhutdinov, Ruslan and Hinton, Geoffrey E.},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2012},
  pages     = {2264-2271},
  doi       = {10.1109/CVPR.2012.6247936},
  url       = {https://mlanthology.org/cvpr/2012/tang2012cvpr-robust/}
}