AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Hyperbolic Neural Operator

Jieyuan Pei, Zhuoxuan Li, Wei Li, Haobo Zhang, jiawei jiang, Jianwei Zheng

ICML 2026 · Architectures and efficiency › Attention and transformers · approach: Transformer neural operators for PDEs

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use hyperbolic distance logits instead of Euclidean logits (ablation)HNO · Darcyrelative L2 error: 0.00446 -> 0.00493 (+10.5%) when removedImproved
add adaptive temperature instead of fixed tau (ablation)HNO · Darcyrelative L2 error: 0.00446 -> 0.00464 (+4.0%) when removedImproved
use hyperbolic distance instead of Euclidean dot-product routing (ablation)HNO · Elasticityrelative L2 error (x1e-3): 3.688 -> 4.100 when removedImproved
use hyperbolic distance instead of Euclidean dot-product routing (ablation)HNO · Airfoil volumerelative L2 error (x1e-3): 12.009 -> 16.683 when removedImproved
use hyperbolic radius instead of Euclidean learned temperature gate (ablation)HNO · Elasticityrelative L2 error (x1e-3): 3.688 -> 5.422 when replaced by TempGateImproved
use hyperbolic radius instead of Euclidean RBF width gate (ablation)HNO · Airfoil volumerelative L2 error (x1e-3): 12.009 -> 133.168 when replaced by RBFGateImproved
keep learned Q/K norms rather than fixing them at inference (ablation)HNO · Darcyrelative L2 error: 0.00446 -> 0.03269 when norm fixedImproved
keep sample-specific Q/K norms rather than shuffling them (ablation)HNO · Darcyrelative L2 error: 0.00446 -> 0.03328 when norm shuffledImproved
apply hyperbolic geometry only to attention logits rather than the full manifold-valued backbone (ablation)HNO · Darcyrelative L2 error: 0.00446 -> 0.00817 with full manifold backboneImproved
use token-mediated compression instead of point-to-grid patchification (ablation)HNO · Darcyrelative L2 error: 0.00446 -> 0.00744 with H-SparsePatchImproved
use HNO instead of Transolver++Transolver++ · Elasticityrelative L2 error: 0.0064 -> 0.0037Improved
use HNO instead of TransolverTransolver · Darcyrelative L2 error and params: 0.00567 -> 0.00446; 2.83M -> 0.82MImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy