🚀 Excited to share our latest research on quantifying the predictive uncertainty of machine learning models. QUAM searches for adversarial models (not adversarial examples!) to better estimate the epistemic uncertainty, the uncertainty about chosen model parameters. 1/5

Interested in reliable prediction intervals for time series? 🚀 Excited to announce our latest work, which proposes HopCPT, was accepted at #Neurips2023 🎉. HopCPT tackles the challenges of time series data for Conformal Prediction with the help of Modern Hopfield Networks. 1/4

Our latest paper challenges the status quo in Unsupervised Learning for Combinatorial Optimization (CO) problems.💡 🔄 Unveiling the limitations of popular Mean Field methods, we showcase the power of Autoregressive Approximations of solving CO Problems on graphs. 🧵

HyperPCM has been published in JCIM🧬 It's the first HyperNetwork dedicated to learning drug-target interactions with enhanced generalization to zero-shot tasks! 📰pubs.acs.org/doi/10.1021/ac… 🤗huggingface.co/spaces/emmas96… Work with Pieter-Jan Hoedt Sepp Hochreiter Günter Klambauer in https://bsky.app/profile/aidd.bsky.social.

VN-EGNN: E(3)-Equivariant GNNs with Virtual Nodes Enhance Protein Binding Site Identification New method to find binding pockets of proteins. Virtual nodes allow to employ distance losses directly. P: arxiv.org/abs/2404.07194 C: github.com/ml-jku/vnegnn 🤗:huggingface.co/spaces/ml-jku/…

xLSTM: Extended Long Short-Term Memory The famous LSTM nets are improved by exponential gates and matrix-memory with covariance update. Strong results on large-scale language modeling. P: arxiv.org/abs/2405.04517

xLSTM Explained in Detail!!! youtu.be/KuRpxvMMrlk?si… via YouTube

ELLIS Machine Learning for Molecules workshop December 6, 2024, HYBRID Paper submission deadline: November 1st. Program chairs: Francesca Grisoni @jtmargraf an me Webpage: moleculediscovery.github.io/workshop2024/ We hope to see you all there!

Keynote at SETAC 2024 was a great opportunity to promote ai-dd.eu & https://bsky.app/profile/aichemist.bsky.social, e.g., contrastive learning by Ana Sanchez-Fernandez/Günter Klambauer, representation learning by Tox24 Challenge or #TransOrGAN published by our associated partner Weida Tong at SI of Chem. Res. Toxicol.

Transformers can be slow for real-time applications like robotics. We study if modern recurrent architectures, like xLSTM and Mamba, can be faster alternatives. Experiments on 432 tasks show that they compare favourably in terms of performance and speed 🎃 arxiv.org/abs/2410.22391

Bio-xLSTM: Generative modeling, representation and in-context learning of biological and chemical sequences xLSTM also shines for DNA, proteins and small molecules -- can handle large-range interactions and huge context! P: arxiv.org/abs/2411.04165

We are excited to introduce Bio-xLSTM! TLDR: we extend xLSTM to genomic, protein and molecular domains and find that it is a proficient generative model, learns rich representations and can perform in-context learning.

The Machine Learning for Molecules workshop 2024 will take place THIS FRIDAY, December 6. Tickets for in-person participation are "SOLD" OUT. We still have a few free tickets for online/virtual participation! Registration link here: moleculediscovery.github.io/workshop2024/

Need to predict bioactivity 🧪 but only have limited data ❌? Try our interactive app for prompting MHNfs — a state-of-the-art model for few-shot molecule–property prediction. No coding or training needed. 🚀 📄 Paper: pubs.acs.org/doi/10.1021/ac… 🖥️ App: huggingface.co/spaces/ml-jku/…

Happy to introduce 🔥LaM-SLidE🔥! We show how trajectories of spatial dynamical systems can be modeled in latent space by --> leveraging IDENTIFIERS. 📚Paper: arxiv.org/abs/2502.12128 💻Code: github.com/ml-jku/LaM-SLi… 📝Blog: ml-jku.github.io/LaM-SLidE/ 1/n

We’re excited to introduce TiRex — a pre-trained time series forecasting model based on an xLSTM architecture.

Ever wondered how linear RNNs like #mLSTM (#xLSTM) or #Mamba can be extended to multiple dimensions? Check out "pLSTM: parallelizable Linear Source Transition Mark networks". #pLSTM works on sequences, images, (directed acyclic) graphs. Paper link: arxiv.org/abs/2506.11997

It's happening again!!! ML4Molecules workshop 2025. within the #ELLIS Unconference, preceding #EurIPS. More infos: moleculediscovery.github.io/workshop2025/