Contrastive Multiview Coding

@article{Tian2019ContrastiveMC,
  title={Contrastive Multiview Coding},
  author={Yonglong Tian and Dilip Krishnan and Phillip Isola},
  journal={ArXiv},
  year={2019},
  volume={abs/1906.05849}
}
Humans view the world through many sensory channels, e.g., the long-wavelength light channel, viewed by the left eye, or the high-frequency vibrations channel, viewed by the right ear. Each view is noisy and incomplete, but important factors, such as physics, geometry, and semantics, tend to be shared between all views (e.g., a ``dog" can be seen, heard, and felt). We hypothesize that a powerful representation is one that models view-invariant factors. Based on this hypothesis, we investigate a… CONTINUE READING
4
Twitter Mentions

Similar Papers

References

Publications referenced by this paper.
SHOWING 1-10 OF 63 REFERENCES

and A

O. J. Hénaff, A. Razavi, C. Doersch, S. Eslami
  • v. d. Oord. Data-efficient image recognition with contrastive predictive coding. arXiv preprint arXiv:1905.09272
  • 2019
VIEW 11 EXCERPTS
HIGHLY INFLUENTIAL

ImageNet: A large-scale hierarchical image database

  • CVPR 2009
  • 2009
VIEW 9 EXCERPTS
HIGHLY INFLUENTIAL

Unsupervised Feature Learning via Non-parametric Instance Discrimination

  • 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition
  • 2018
VIEW 5 EXCERPTS
HIGHLY INFLUENTIAL