University of Luxembourg

4 related articles

Subscribe to updates

arXiv Nov 4, 2026
SO3LR Tauri stretches one force field to 85 elements: a single set of weights runs proteins, nucleic acids, membranes and platinated DNA
SO3LR Tauri extends the earlier SO3LR force field from eight elements and 4 million structures to 85 elements and roughly 148 million structures, and rewrites its short-range repulsion, electrostatics and dispersion modules. Keeping the fast SO3krates message-passing backbone, it reproduces nucleobase-pair and protein–ligand interactions and carbohydrate conformer ordering against coupled-cluster references, and yields water and organic-solvent densities, ion hydration and structures of several biomolecular classes in dynamics without system-specific fitting. The models ship in three sizes and scale to the million-atom range on a single GPU.
SO3LR Tauri extends the earlier SO3LR force field from eight elements and 4 million structures to 85 elements and roughly 148 million structures, and rewrites its short-range repulsion, electrostatics and dispersion modules. Keeping the fast SO3krates message-passing backbone, it reproduces nucleobase-pair and protein–ligand interactions and carbohydrate conformer ordering against coupled-cluster references, and yields water and organic-solvent densities, ion hydration and structures of several biomolecular classes in dynamics without system-specific fitting. The models ship in three sizes and scale to the million-atom range on a single GPU.
University of Luxembourg Technische Universität Berlin
Adil Kabylda（First author）
SO3LR Tauri Machine Learning Interatomic Potentials SO3krates
IEEE Journal on Selected Areas in Communications Oct 9, 2026
GREQ two-phase hybrid framework delivers up to 40% throughput gain and 100% QoS satisfaction in distributed 6G resource management
The work introduces GREQ, a two-phase hybrid model-learning framework: phase one uses a channel-statistics-based greedy algorithm to deterministically satisfy user QoS with minimal subchannels, and phase two exploits the OFDM property that inter-cell interference on a subchannel comes only from the same subchannel to decompose the large-scale MARL problem into independent per-subchannel QMIX subproblems, avoiding exponential action-space growth; in simulations with 50MHz bandwidth and different numerologies, GREQ improves throughput by up to 40% over optimization and learning baselines, achieves 100% constraint satisfaction, and activates only 20–80% of available subchannels.
The work introduces GREQ, a two-phase hybrid model-learning framework: phase one uses a channel-statistics-based greedy algorithm to deterministically satisfy user QoS with minimal subchannels, and phase two exploits the OFDM property that inter-cell interference on a subchannel comes only from the same subchannel to decompose the large-scale MARL problem into independent per-subchannel QMIX subproblems, avoiding exponential action-space growth; in simulations with 50MHz bandwidth and different numerologies, GREQ improves throughput by up to 40% over optimization and learning baselines, achieves 100% constraint satisfaction, and activates only 20–80% of available subchannels.
National University of Singapore University of Luxembourg
Thang X. Vu（First author）
6G Network QMIX Multi-Agent Reinforcement Learning Orthogonal Frequency Division Multiplexing Reinforcement Learning
arXiv Oct 1, 2026
Keeping only the top-attention tokens without retraining, nine checkpoints need effective attention sets of 7 to 283 tokens, and longer context raises the required set size
Without retraining, this work retains only the tokens with the highest attention weights at each head, layer, and query while keeping their original weights unchanged, and estimates the effective attention set size needed to stay within a chosen loss tolerance by measuring the increase in negative log-likelihood (NLL); it finds that relatively small selected sets keep NLL close to the full-attention baseline, that attention-based selection substantially outperforms random selection, that the required set size grows with context while its fraction of context decreases, that in BABILong experiments with a fixed annotated supporting fact additional background pushes support tokens down the attention ranking and reduces their attention mass, and that renormalizing the retained weights can subs
Without retraining, this work retains only the tokens with the highest attention weights at each head, layer, and query while keeping their original weights unchanged, and estimates the effective attention set size needed to stay within a chosen loss tolerance by measuring the increase in negative log-likelihood (NLL); it finds that relatively small selected sets keep NLL close to the full-attention baseline, that attention-based selection substantially outperforms random selection, that the required set size grows with context while its fraction of context decreases, that in BABILong experiments with a fixed annotated supporting fact additional background pushes support tokens down the attention ranking and reduces their attention mass, and that renormalizing the retained weights can subs
University of Luxembourg London Institute for Mathematical Sciences
Timur Mudarisov（First author）
Self-Attention Token Selection Language modeling Question Answering
arXiv Oct 1, 2026
Across six pretrained language models, residual streams favor their own endpoint early while competitor sets shrink with changing membership
By comparing each context's intermediate residual states with its own final state and an empirical bank of final states from other contexts, the study finds across Gemma-2B, Gemma-7B, Qwen2.5-1.5B, Qwen2.5-7B, Mistral-7B, and Llama-3-8B that the own endpoint becomes preferable to the average alternative at the earliest measured layer while many individual endpoints remain closer, that competitor sets generally shrink with depth yet show both entries and exits, and that directional alignment can improve while Euclidean distance to the final state changes little; the authors add a high-dimensional model separating norm, alignment, and endpoint geometry and prove that a straight path toward the own endpoint cannot introduce new competitors.
By comparing each context's intermediate residual states with its own final state and an empirical bank of final states from other contexts, the study finds across Gemma-2B, Gemma-7B, Qwen2.5-1.5B, Qwen2.5-7B, Mistral-7B, and Llama-3-8B that the own endpoint becomes preferable to the average alternative at the earliest measured layer while many individual endpoints remain closer, that competitor sets generally shrink with depth yet show both entries and exits, and that directional alignment can improve while Euclidean distance to the final state changes little; the authors add a high-dimensional model separating norm, alignment, and endpoint geometry and prove that a straight path toward the own endpoint cannot introduce new competitors.
University of Luxembourg London Institute for Mathematical Sciences
Timur Mudarisov（First author）
Geometric Inference Hypothesis Transformer Residual Streams Endpoint Bank Projected Normal Distribution von Mises–Fisher Distribution

You’re all caught up

SO3LR Tauri stretches one force field to 85 elements: a single set of weights runs proteins, nucleic acids, membranes and platinated DNA

GREQ two-phase hybrid framework delivers up to 40% throughput gain and 100% QoS satisfaction in distributed 6G resource management

Keeping only the top-attention tokens without retraining, nine checkpoints need effective attention sets of 7 to 283 tokens, and longer context raises the required set size

Across six pretrained language models, residual streams favor their own endpoint early while competitor sets shrink with changing membership