Federico Monti

Representation Learning on graphs. Ex. Sr. ML Research Scientist @TwitterUK, former founder and CTO @Fabula_ai, addicted traveller.

[Michael Bronstein @mmbronstein]

A new @TDataScience blog post with @emaros96
discusses network games and how to learn unknown interactions of players from their outcomes:

https://t.co/r8W7r5qzac

More details in the ICML 2022 paper with @frederickmonti
@lylengyan Xiaowen Dong:

https://t.co/8105c53Cvx

[Emanuele Rossi @emaros96]

At ICML? Join the presentation of our recent paper "Learning to Infer Structures of Network Games" by @lylengyan today (Wed) at 9:35 local time, game theory track!
https://t.co/alN9m75IcY
@mmbronstein @epomqo @frederickmonti

[Emanuele Rossi @emaros96]

Many interactions between individuals can be modeled as network games. Given the actions taken by individuals, can we recover the underlying network structure?

In a new ICML paper combining game theory, graphs and machine learning, we show that you can!
https://t.co/alN9m75IcY

[DeepGraphLibrary @DGLGraph]

[3/5]

- SIGN Diffusion Transform by @ffabffrasca, @emaros96, @aittalam, @b_p_chamberlain, @mmbronstein, and @frederickmonti

- Label Propagation by Xiaojin Zhu and @ZoubinGhahrama1

[Ana-Maria Cretu @AnaMariaCretu5]

I'll be presenting my work on profiling attacks against interaction data later today at the Privacy Preserving Data Analysys Workshop. Looking forward to it!
If you want to register, here is the link https://t.co/hi0GIhqYXU

[Federico Monti @frederickmonti]

Honoured to have been selected as “highlighted reviewer” among few others for @iclr_conf!

Looking forward to attending the conference next week!

[Yves-A. de Montjoye @yvesalexandre]

🚨 New profiling attack in @NatureComms: we show, using graph neural networks, how interaction data such as messages or bluetooth close proximity metadata can be used to uniquely identify individuals over long periods of time. https://t.co/U5xtXaLM0O A thread 🧵

[Luca (also on the other platforms) @__lucab]

#ItsComingRome 🇮🇹🇪🇺

[Federico Monti @frederickmonti]

@aittalam @ffabffrasca For sure more interpretable and easy to use than Scalding 😂

[Federico Monti @frederickmonti]

@aittalam @ffabffrasca facimm na' strage mmiez 'a Source e rifacimm accussì!

[Emanuele Rossi @emaros96]

Our model, SIGN (https://t.co/DQqjV929h3) obtains SOTA results on OGB-papers100M, the largest public node-classification dataset with over 110M nodes! An epoch takes less than a minute! 🎉🎉🎉
Joint work w/ @ffabffrasca @mmbronstein @frederickmonti @b_p_chamberlain @aittalam

[Nils Hammerla @nhammerla]

We’re looking to hire a Machine Learning Research Engineer to join the Learning Methods Research Team. In this role, you’ll take the latest ideas from research to production and make them scale! Let me know if you are interested or apply here https://t.co/sI2Vjnz4Pp

[Emanuele Rossi @emaros96]

The code for our recent paper "TGN: Temporal Graph Networks" is live on Github: https://t.co/5NZPOZnPJ9 !
Paper link: https://t.co/1gUTxlC3pr
Joint work w/ @mmbronstein @frederickmonti @b_p_chamberlain @ffabffrasca @aittalam

[Michael Bronstein @mmbronstein]

New blog post on temporal graph networks for deep learning on dynamic graphs written with @emaros96 Our work with Twitter team @frederickmonti @b_p_chamberlain @ffabffrasca @aittalam

[Petar Veličković @PetarV_93]

1 hour until the #ICML2020 Workshop on Graph Representation Learning and Beyond starts!

Tune in all-day long: https://t.co/kHX85WMfXl

@mmbronstein @andreeadeac22 @williamleif @jhamrick @miladhash @jure @lrjconan @frederickmonti @YizhouSun @kswersk @RexYing0923 @marinkazitnik

[Federico Monti @frederickmonti]

In need of a solution for inference on temporal graphs? Check out our #neurips2020 submission! https://t.co/m3VrNmJa1D Via a combination of memory modules and graph conv layers we achieve SOTA results on a variety of benchmarks while being more efficient wrt previous solutions!

[Emanuele Rossi @emaros96]

Our new GNN model for large graphs, SIGN, is already available on Pytorch Geometric only days after we put in on arxiv! Shout-out to the creator of pytorch geometric @rusty1s for his work

https://t.co/OY18nj6QTI

https://t.co/DQqjV929h3

[Emanuele Rossi @emaros96]

New work on scalable GNNs: SIGN scales to very large graphs thanks to efficient precomputation of inception-like graph filters! We are competitive with the SOTA while being 10x faster! https://t.co/DQqjV929h3 @ffabffrasca @frederickmonti @mmbronstein @aittalam @b_p_chamberlain

[Federico Monti @frederickmonti]

@thomaskipf @wellingmax Congrats Thomas!!

[Petar Veličković @PetarV_93]

Our GRL+ Workshop at @icmlconf #ICML2020 now directly encourages submissions that use graph representation learning to mitigate impacts of #COVID19 (3 contributed talks)!

We prepared resources that outline relevant GRL tasks and datasets at https://t.co/jX9FWlXWGd.