Learning the Structure of Graph Neural Networks | Mathias Niepert | heidelberg.ai

Heidelberg AI Talk 9th July 2019 | Learning the Structure of Graph Neural Networks | Mathias Niepert, NEC Labs Europe
Related papers:
arxiv.org/abs/1903.11960
heidelberg.ai

[@jackflash9123]

Thank you for uploading this talk! Really high quality and clear explanations by the speaker, super nice :)

[@TheAnna1101]

Great talk, rich content and a lot of information. Thank you!

[@Uvideo4tube]

Brilliant Talk . Very informative and lucid explanation . Please share the ppt link .

[@gabrielemazzola9652]

Thanks for the awesome video, the explanation is very clear.
Just a small note: at 22:34, you mention the "dot product" operation, but this is not a dot product. Instead, it's a simple vector-matrix multiplication.

[@mniepert]

Thanks for the feedback! Glad you liked it. I’m very self-conscious about talks and always wonder if I actually did a decent job.
You are right by the way, the first operation is a vector-matrix multiplication (several dot products) and the second one a single dot product.

[@user-sm1wy6qz1v]

Thank you for brilliant talk. How can I download the talk materials? I want to download the ppt as pdf or anything else.

[@mniepert]

Thanks for your nice feedback. Here is the pdf: matlog.net/hdai.pdf

[@sirisaksirisak6981]

Specify to layer of layers of node connection to all connection.Frame all total network.

[@cybervigilante]

Bell Labs used to support pure research. Now, the only high level support of pure research in the US is from the military. We are stagnating.