Saturday, November 2, 2024
Latest:
Videos

Geometric Deep Learning

Artis Modus



Siraj Raval

Geometric Deep Learning is able to draw insights from graph data. That includes social networks, sensor networks, the entire Internet, and even 3D Objects (if we consider point cloud data to be a graph). I’ll explain how it works via a demo of me using a graph convolutional network to classify people by their interest in sports teams as well as a 3D object classification demo. At its core, it comes down to being able to learn from non-Euclidean data. Euclid’s laws help define certain types of data, so I’ll cover some geometry background as well. Enjoy!

Code for this video:
https://github.com/llSourcell/pytorch_geometric

Please Subscribe! And Like. And comment. Thats what keeps me going.

Want more education? Connect with me here:
Twitter: https://twitter.com/sirajraval
instagram: https://www.instagram.com/sirajraval
Facebook: https://www.facebook.com/sirajology

More learning resources:
http://sungsoo.github.io/2018/02/01/geometric-deep-learning.html
http://geometricdeeplearning.com/
https://arxiv.org/abs/1611.08097
http://3ddl.stanford.edu/CVPR17_Tutorial_Intrinsic_CNNs_compressed.pdf
https://github.com/rusty1s/pytorch_geometric
https://pemami4911.github.io/paper-summaries/deep-learning-theory/2017/11/19/geometric-deep-learning.html

Join us at the School of AI:
https://theschool.ai/

Join us in the Wizards Slack channel:
http://wizards.herokuapp.com/

Please support me on Patreon:
https://www.patreon.com/user?u=3191693

Signup for my newsletter for exciting updates in the field of AI:
https://goo.gl/FZzJ5w

Hiring? Need a Job? See our job board!:
https://www.theschool.ai/jobs/

Need help on a project? See our consulting group:
https://www.theschool.ai/consulting-group/

Source

Similar Posts

23 thoughts on “Geometric Deep Learning”

  1. This would be a great alternative to CNN, RNN and LSTM text classifiers that use word2vec. Word2vec is a three dimensions represents of words, so these new geometric models could be the next step in natural language processing.

  6. Has there been any attempt to do pointcloud segmentation applying this kind of network layers in a similar manner to the U-Net for 2D? Anyone up for the challenge?

  8. Hey Siraj,

    Could you do a video explaining spiking neural networks and the use of them, as the search results for it on YouTube are not very satisfactory.

    Keep it up with your videos, awesome stuff!

  11. Very very interesting video! Keep up Siraj! I have one question regarding this. How do these methods differ from more traditional Probabilistic Graphical Modeling with missing data? Most of these problems seem manageable through some typical Belief/Markov metwork afaik. For example, given a graphical independence structure it is trivial (at least in theory maybe not so for large N) to predict MLE/MAP of missing verticies. Is this approach different due to efficiency or is there more theoretical poterntial?

  16. Perhaps I missunderstood, but Euklidian space is not just 1 or 2D but also includes higher dimensions R^n. You can describe a sphere in Euklidian space like r^2 = x^2 + y^2 + z^2. Euklidian space is based on the notion of uncurved space.

  17. Hi Siraj,
    I don´t know if it is a silly question but it came to my mind: Is it possible to combine Geometric Deep Learning with Hinton´s capsule nets? Because the vectors in each capsule can be three- or n-dimensional…

  20. Maybe because it's late, I did not grasp the non-euclidean convolution stuff along with receptive field. Maybe you can point to some other refs, or make a new vid ? 😉

  21. Im REALLY interested in 3d object recognition and pose estimation, preferably from 3d Mesh than point cloud but point cloud works as well….. Could you do a video of implementation of GCNs? That would be AWESOME!
    THANKS for always keeping us up to date with the latest in AI!!

Comments are closed.

WP2Social Auto Publish Powered By : XYZScripts.com