Lightnews — Scholar-powered news

Lev Telyatnikov @levtelyatnikov.bsky.social · Aug 25

We can’t wait to see what the community builds with TopoBench! Star our repo, read the paper, and bring reproducibility to your TDL research. Contributions and feedback are welcome! #OpenSource #Python #AcademicTwitter

4 6

Lev Telyatnikov @levtelyatnikov.bsky.social · Aug 25

This was a massive team effort!
Developed & led by @levtelyatnikov.bsky.social , @gbg141.bsky.social , Theodore Papamarkou. With major contributions from @Marco Montagna, Mustafa Hajij, Ghada Zamzmi, Michael T. Schaub, @ninamiolane.bsky.social , & @sscardapane.bsky.social , and many others!

1 4 6

Lev Telyatnikov @levtelyatnikov.bsky.social · Aug 25

For a deep dive into our framework for reproducible benchmarking, our methodology, and key findings, check out the full paper. Paper: arxiv.org/abs/246.06642 #MachineLearning #GNN #DataScience

arxiv.org

1 3 5

Lev Telyatnikov @levtelyatnikov.bsky.social · Aug 25

At its core, TopoBench is a framework for robust and fair evaluation in TDL. We’re empowering researchers and practitioners to compare models with confidence and push the field forward, together.

1 2 4

Lev Telyatnikov @levtelyatnikov.bsky.social · Aug 25

Curious about topological liftings? TopoBench lets you experiment with them on your own data! We’ve included 20+ liftings from last year’s TDL Challenge @TAGinDS, @PyT_Team_. Explore them all in our Wiki: github.com/geometric-in... #TDA

Structural Liftings

TopoBench is a Python library designed to standardize benchmarking and accelerate research in Topological Deep Learning - geometric-intelligence/TopoBench

github.com

1 3 3

Lev Telyatnikov @levtelyatnikov.bsky.social · Aug 25

Struggling with reproducible benchmarks in TDL? We’ve got you covered.

12+ models tested on 22+ diverse, cross-domain datasets. Tutorials to get you started with topology & higher-order homophily Webpage/Docs: geometric-intelligence.github.io/TopoBench/

TopoBench

TopoBench: A comprehensive benchmarking framework for Topological Deep Learning models and datasets.

geometric-intelligence.github.io

1 3 3

Lev Telyatnikov @levtelyatnikov.bsky.social · Aug 25

Delighted to announce TopoBench has been accepted to DMLR! It’s a modular library for Topological Deep Learning, built to provide reproducible, cross-domain benchmarks and accelerate research. GitHub: github.com/geometric-in... #AI #DeepLearning

GitHub - geometric-intelligence/TopoBench: TopoBench is a Python library designed to standardize benchmarking and accelerate research in Topological Deep Learning

TopoBench is a Python library designed to standardize benchmarking and accelerate research in Topological Deep Learning - geometric-intelligence/TopoBench

github.com

1 4 13

Reposted by Lev Telyatnikov

martinca.bsky.social @martinca.bsky.social · May 26

TDL actually works at scale! And we believe 𝐇𝐎𝐏𝐒𝐄 lays the foundation for broad applications of TDL ✨

📭 Reach out for collaborations

Special thanks to @levtelyatnikov.bsky.social and @gbg1441 and the team @ninamiolane.bsky.social and @Coerulatus :)
(6/6)

5 7

Reposted by Lev Telyatnikov

Nina Miolane @ninamiolane.bsky.social · May 27

At last Topological Neural Networks are fast🚀

HOPSE builds an encoder for combinatorial complexes, enabling topological deep learning (TDL) w/o the usual computational cost.

A major step forward for TDL!
@martinca.bsky.social @gbg141.bsky.social @marcomonga.bsky.social @levtelyatnikov.bsky.social

martinca.bsky.social @martinca.bsky.social · May 26

🚨Higher-order combinatorial models in TDL are notoriously slow and resource-hungry. Can we do better?

Introducing:
🚀 𝐇𝐎𝐏𝐒𝐄: A Scalable Higher-Order Positional and Structural Encoder for Combinatorial Representations 🚀

📝 arXiv: arxiv.org/abs/2505.15405

🧵 (1/6)

4 7

Reposted by Lev Telyatnikov

Geometric Intelligence Lab @geometric-intel.bsky.social · May 27

🚨 New preprint from the lab!

Discover fast topological neural networks, that leverage higher order structures without the usual computational burden!

By @martinca.bsky.social @gbg141.bsky.social @marcomonga.bsky.social @levtelyatnikov.bsky.social @ninamiolane.bsky.social

2 2

Lev Telyatnikov @levtelyatnikov.bsky.social · May 26

Absolutely proud of this work! Huge thanks to @gbg141.bsky.social @ninamiolane.bsky.social @marcomonga.bsky.social — and of course @martinca.bsky.social, who drove the project, learned on the fly, and kept the enthusiasm high at every turn!

martinca.bsky.social @martinca.bsky.social · May 26

🚨Higher-order combinatorial models in TDL are notoriously slow and resource-hungry. Can we do better?

Introducing:
🚀 𝐇𝐎𝐏𝐒𝐄: A Scalable Higher-Order Positional and Structural Encoder for Combinatorial Representations 🚀

📝 arXiv: arxiv.org/abs/2505.15405

🧵 (1/6)

1 5 8