🔥Telegram İfşa

I'm looking forward to presenting our new paper "Better Set Representations For Relational Reasoning" tonight at #NeurIPS2020: arxiv.org/abs/2003.04448! This is joint work with another first author Horace He, also Abhay Singh, Yan Zhang, Ser Nam, and Austin Benson. (1/n)

Come chat with us about set prediction, permutation equivariance, implicit differentiation, and much more! We will be at spot E3 room 1 in gather town today at 7:30pm-9:30pm (UTC+2). Link: iclr.cc/virtual/2022/p… Joint work with Yan Simon Lacoste-Julien Gertjan Burghouts Cees Snoek

I visited Japan last week. It was amazing!

SAIL Montreal (SAIT AI Lab Montreal from Samsung) is happy to sponsor #NeurIPS2022 this week! See sail-montreal.github.io/neurips2022/ for our 5 NeurIPS papers; check out our booth at Expo Hall #224 and meet our research scientists! Note: we're still hiring!

New #NeurIPS2022 workshop paper: we improve Slot Attention by using Optimal Transport ideas to make it *multiset*-equivariant (not permutation/set-equivariant). Poster from 1-2pm and 4-5pm CST at the AllThingsAttention workshop in Room 399 today! attention-learning-workshop.github.io/2022/papers/zh…

Want to make your model equivariant to certain known transformations without changing the network architecture? Please attend our contributed talk at 11AM Symmetry and Geometry in Neural Representations to know how we achieve this by adding a simple shallow network that learns to canonicalize the input. #NeurIPS2022

In this new joint work (Emy Gervais, @FatrasKilian, Yan, Simon Lacoste-Julien), we show the massive power of averaging the weights of multiple models trained simultaneously! 🤯 Blog: ajolicoeur.wordpress.com/papa Website: github.com/SamsungSAILMon… Arxiv: arxiv.org/abs/2304.03094

SAIL Montreal (SAIT AI Lab Montreal from Samsung) will have a strong presence at #icml2023 next week with 8 ICML papers! See sail-montreal.github.io for the list and we'll have 6 research scientists attending who will be happy to chat with you...

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!