Deep Graph Networks

Did you know that you can build expressive and efficient equivariant message passing for molecular systems using Cartesian rather than Spherical tensors (e.g., MACE)? If not, it’s time to have a look at our NeurIPS 24 paper led by Viktor!

Graph-Induced Sum-Product Networks have been (finally!) accepted at ICLR 2024! Blog post coming in the following months… stay tuned! Special thanks to Mathias Niepert for his relentless support =).

We propose a probabilistic model that bridges graph machine learning and sum-product networks to tractably answer probabilistic queries. Our Graph-induced Sum-Product Network can solve unsupervised and supervised graph tasks and it is especially effective in modeling the missing data distribution as well as exploiting unlabeled data in a scarce supervision scenario.

PyDGN just got accepted in Journal of Open Source Software =) I have to say, compared to other prestigious journals, I got much more constructive feedback and the review process was faster and transparent.

I am really excited: my very first single-author paper has been accepted at NeurIPS 2023! I still can’t believe that I managed to pull this off, especially after submitting it for the first time =) give it a few weeks for the papers to be released on OpenReview!

Infinite Contextual Graph Markov Model

PyDGN 1.0.0 is out! We have added a documentation, refactored the code, implemented a bunch of user-friendly features and provided clean, well commented configuration files!

PyDGN v0.5.0 is out

Graph Mixture Density Network

Deep Bayesian Graph Networks & Low-resource Classification