AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Positional Encoding Is All You Need For Scalable Equivariance Constraint Relaxation

Hagay Michaeli, Haggai Maron, Daniel Soudry

NeurIPS 2026 · Architectures and efficiency › RNNs and normalization · approach: Symmetry and equivariance in loss landscapes · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes using positional encodings to relax equivariance constraints in neural networks, allowing approximate symmetry while retaining scalability (details limited to the abstract).

Problem

Strict equivariance can hinder optimization when data only approximately satisfy the assumed symmetry or when equivariance constraints create a difficult loss landscape.

Approach

Relaxes strict equivariance constraints via positional encodings rather than fully equivariant architectures; the abstract does not give further specifics.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add annealed positional-encoding relaxation (additive)VN-PointNet · ModelNet40 SO(3)/SO(3)instance accuracy: 67.5 -> 79.6Improved
use PE relaxation instead of dense-layer CRVN-PointNet · ModelNet40 SO(3)/SO(3)instance accuracy: 74.0 -> 79.6Improved
add annealed PE relaxationVN-DGCNN · ModelNet40 SO(3)/SO(3)instance accuracy: 88.3 -> 88.6No clear effect
use PE instead of dense relaxationVN-PointNet · ModelNet40 SO(3)/SO(3)training time per epoch (s): 97.6 -> 27.9Improved
add annealed PE relaxationRapidash-Eq · ModelNet40 Rapidashaccuracy: 83.7 -> 84.6Improved
add annealed PE (train only) instead of keeping PE at testAF-XCiT-Nano · ImageNet shift accuracyshift accuracy: 70.3 -> 71.1Improved
add annealed PE relaxationAF-XCiT-Nano · ImageNet valtop-1 accuracy: 70.6 -> 71.1Improved
add annealed PE relaxationE2CNN · Smoke plume (Future)RMSE x1e-2: 81.2 -> 79.8Improved
add PE relaxationSEGNN · N-Bodytest MSE x1e-4: 44.3 -> 41.9Improved
make PE scale gamma learnable (ablation)VN-PointNet · ModelNet40 SO(3)/SO(3) ablationaccuracy w/o PE: 4.5 -> 79.7Improved
add linear decay schedule with learnable gamma (ablation)VN-PointNet · ModelNet40 SO(3)/SO(3) ablationaccuracy w/o PE: 79.7 -> 80.4No clear effect
add explicit regularization beta=0.001 (learnable gamma) (ablation)VN-PointNet · ModelNet40 SO(3)/SO(3) ablationaccuracy w/o PE: 79.7 -> 80.7No clear effect

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