Aachen Machine Learning Meetup

Dec 5, 2017 · Aachen, Germany

We will quickly discuss a practical example of word embeddings and then introduce to "capsule networks".

Capsule networks (CapsNet) introduce a new tool in deep learning to better model hierarchical relationships by nesting sets of neural layers. Sabour, Frosst and Hinton (2017) show that "a discrimininatively trained, multi-layer capsule system achieves state-of-the-art performance on MNIST and is considerably better than a convolutional net at recognizing highly overlapping digits". According to Anonymous (2017), benchmarks show that capsule networks are also more resistant to white box adversarial attacks than convolutional neural networks (CNN).

Co-creator: Geoffrey Hinton

Background:

Hinton, G. E., Krizhevsky, A., & Wang, S. D. (2011, June). Transforming auto-encoders. In International Conference on Artificial Neural Networks (pp. 44-51). Springer Berlin Heidelberg. Available at:  http://www.cs.toronto.edu/~fritz/absps/transauto6.pdf

Sabour, S., Frosst, N., & Hinton, G. E. (2017). Dynamic Routing Between Capsules. arXiv preprint arXiv:[masked]. Available at https://arxiv.org/pdf/1710.09829.pdf

Anonymous (2017). Matrix capsules with EM routing. In International Conference on Learning Representations 2018 (ICLR 2018). Available at: https://openreview.net/pdf?id=HJWLfGWRb

Event organizers
  • Aachen Machine Learning Meetup

    This is an interest-group for practitioners and researchers at RWTH Aachen and beyond, working on machine learning (ML).  In our meetings, we present and discuss most recent research advances and showcase examples of implementation. By doing so, we address both the scientific community, as well as startup founders and companies that like to present their innovations and job opportunities. All skill levels are welcome. You like to present your research, startup or company? Send an e-mail to the organizers a

    Recent Events
    More

Are you organizing Aachen Machine Learning Meetup?

Claim the event and start manage its content.

I am the organizer
Social
Rating

based on 0 reviews