#cg/memo

"Graph Neural Networks can Recover the Hidden Features Solely from the Graph Structure" が #ICML2023 に採択

twitter.com/joisino_/sta...

arxiv.org/abs/2301.10956

We derive a specific instance of encoder-decoder models, the Coupled Variational Autoencoder (#ICML2023 arxiv.org/abs/2306.02565) as implementing cooperative communication over imperfect common ground.

Just came across a blog post from Hussein Hazimeh about how Google has extended the #icml2022 paper in which @Xiny_win and I worked closely together in the paper that their team published at #icml2023:

blog.research.google/2023/08/neural…

It is always great to see our work used in practice!

How can we effectively use neural networks to process the parameters of other neural nets?

We propose representing neural networks as computation graphs, enabling the use of standard graph neural networks to preserve permutation symmetries.
Join our poster @TAGinDS in #ICML2023!

Michael Hu is presenting our work at the High Dimensional Learning Dynamics workshop at #ICML2023. It has the coolest visualizations I’ve ever touched. It’s the Mona Lisa of training dynamics posters. There’s gonna be a corner of the gallery full of weeping empiricists mumbling “only art is real”.

Just re-upping this. Lots of Jane Street folk at #ICML2023. You can come by our booth there to say hello. (It's full of some really fun puzzles if you like that sort of thing...)

Today at #ICML2023 I will be co-presenting "Graph Switching Dynamical Systems". Join us in poster session 2, poster 636.

Joint work with Yongtuo Liu, @saramagliacane and @egavves.

Arxiv:Github:

github.com/yongtuoliu/Gra… arxiv.org/abs/2306.00370

Excited to be in Hawaii for #ICML2023! Ping me if you want to grab a coffee and chat about graph networks, geometric deep learning, and neural fields.

Just a heads up that there's a ton of Jane Street folk at #ICML2023, so keep an eye out for them if you want to hear a bit more about what we're up to in the ML space.

http://janestreet.com/icml

in Honolulu for #ICML2023, dm if you want to hang out :)