Michael Galkin

SymmCD: Symmetry-Preserving Crystal Generation with Diffusion Models Generating novel crystalline materials has potential to lead to advancements in fields such as electronics, energy storage, and catalysis. The defining characteristic of crystals is their symmetry,...

2. SymmCD: Symmetry-Preserving Crystal Generation with Diffusion Models

openreview.net/forum?id=xns...

With Daniel Levy, @sibasmarak.bsky.social, @oumarkaba.bsky.social, Santiago Miret, Kelvin Lee, Qiang Zhu, and and Siamak Ravanbakhsh

Fully-inductive Node Classification on Arbitrary Graphs One fundamental challenge in graph machine learning is generalizing to new graphs. Many existing methods following the inductive setup can generalize to test graphs with new structures, but...

1. GraphAny: Fully-inductive Node Classification on Arbitrary Graphs
openreview.net/forum?id=1Qp...

With Jianan Zhao, @zhaochengzhu.bsky.social, @mmbronstein.bsky.social, and @tangjian.bsky.social

✈️ I'll be at ICLR in Singapore and present two papers (happened to be at the same Thu 24th afternoon poster session, some quantum entanglement will be required 👀).
Happy to chat about graph learning and stuff ☕
Paper links in 🧵

*Please repost* @sjgreenwood.bsky.social and I just launched a new personalized feed (*please pin*) that we hope will become a "must use" for #academicsky. The feed shows posts about papers filtered by *your* follower network. It's become my default Bluesky experience bsky.app/profile/pape...

Excellent work led by @hxyscott.bsky.social on expressiveness of KG foundation models - the road to higher-order FMs is now open! 📈

SymmCD: Symmetry-Preserving Crystal Generation with Diffusion Models Generating novel crystalline materials has potential to lead to advancements in fields such as electronics, energy storage, and catalysis. The defining characteristic of crystals is their symmetry,...

2/2 SymmCD: Symmetry-Preserving Crystal Generation with Diffusion Models
openreview.net/forum?id=xns...

One of the coolest inspirations of cutting winter snowflakes ❄️
Led by Daniel Levy, Siba Smarak Panigrahi, @oumarkaba.bsky.social, Qiang Zhu, Kelvin Lee, Santiago Miret, and Siamak Ravanbakhsh

Fully-inductive Node Classification on Arbitrary Graphs One fundamental challenge in graph machine learning is generalizing to new graphs. Many existing methods following the inductive setup can generalize to test graphs with new structures, but...

1/2 Fully-inductive Node Classification on Arbitrary Graphs
openreview.net/forum?id=1Qp...

GraphAny, a zero-shot inductive node classifier for any graph for any input features! Led by Jianan Zhao, Hesham Mostafa, @zhaochengzhu.bsky.social @mmbronstein.bsky.social and @tangjian.bsky.social

Celebrating 2 papers accepted at ICLR 2025 advancing generalist GNNs and generative models for materials discovery 🎉 .
Thread below 🧵

Come say hi to us at NeurIPS 👋
* Tue: AI 4 EDA workshop
* Wed: Temporal Graph Benchmark 2.0, poster #3000 East A-C, 4:30-7:30pm
* Thu: Ultra Deluxe 💅 (graph foundation models): poster #3011, East A-C, 11am-2pm
* Sat: GraphAny poster at the Adaptive Foundation Models workshop

Carrying the theatrical "ice bucket challenge" by recent posts of @mmbronstein.bsky.social and @petar-v.bsky.social - we attended Il Trovatore at the Metropolitan Opera!
Now passing the torch to the renowned opera connoisseur @chrsmrrs.bsky.social

Yes please!

Hello 🦋! I joined GoogleAI as a Senior Research Scientist and moved to NYC, looking for tips on how to handle East Coast winters after eternal summers of SoCal (MetOpera and The Nutcracker ballet are already booked). Also going to be at NeurIPS in Vancouver next week!

Please add me :)