Saved in:
Bibliographic Details
Main Authors: Liu, Zhaoyi, Zhang, Haichuan, Li, Ang
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.20896
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910062663958528
author Liu, Zhaoyi
Zhang, Haichuan
Li, Ang
author_facet Liu, Zhaoyi
Zhang, Haichuan
Li, Ang
contents Hyper-Connections (HC) generalize residual connections into multiple streams, employing residual matrices for cross-stream feature mixing to enrich model expressivity. However, unconstrained mixing disrupts the identity mapping property intrinsic to the residual connection, causing unstable training. To address this, Manifold-Constrained Hyper-Connections (mHC) and its variant restrict these matrices to the Birkhoff polytope (doubly stochastic matrices) via Sinkhorn iterations or permutation-based parameterizations. We reveal three limitations of this polytope constraint: (1) identity degeneration, where learned matrices collapse around the identity and diminish cross-stream interactions, (2) an expressivity bottleneck, as the non-negativity constraint prevents subtractive feature disentanglement, and (3) parameterization inefficiencies, manifesting as unstable Sinkhorn iterations or the factorial-scaling overhead of permutation-based parameterizations. To overcome these flaws, we propose Spectral-Sphere-Constrained Hyper-Connections (sHC). By geometrically shifting the feasible set from a rigid polytope to a spectral norm sphere, sHC allows negative entries, unlocking subtractive interactions for selective feature diversification. This shift eliminates unstable Sinkhorn projections and factorial parameterization, enabling expressive, non-degenerate residual matrices while preserving training stability.
format Preprint
id arxiv_https___arxiv_org_abs_2603_20896
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Beyond the Birkhoff Polytope: Spectral-Sphere-Constrained Hyper-Connections
Liu, Zhaoyi
Zhang, Haichuan
Li, Ang
Machine Learning
Artificial Intelligence
Hyper-Connections (HC) generalize residual connections into multiple streams, employing residual matrices for cross-stream feature mixing to enrich model expressivity. However, unconstrained mixing disrupts the identity mapping property intrinsic to the residual connection, causing unstable training. To address this, Manifold-Constrained Hyper-Connections (mHC) and its variant restrict these matrices to the Birkhoff polytope (doubly stochastic matrices) via Sinkhorn iterations or permutation-based parameterizations. We reveal three limitations of this polytope constraint: (1) identity degeneration, where learned matrices collapse around the identity and diminish cross-stream interactions, (2) an expressivity bottleneck, as the non-negativity constraint prevents subtractive feature disentanglement, and (3) parameterization inefficiencies, manifesting as unstable Sinkhorn iterations or the factorial-scaling overhead of permutation-based parameterizations. To overcome these flaws, we propose Spectral-Sphere-Constrained Hyper-Connections (sHC). By geometrically shifting the feasible set from a rigid polytope to a spectral norm sphere, sHC allows negative entries, unlocking subtractive interactions for selective feature diversification. This shift eliminates unstable Sinkhorn projections and factorial parameterization, enabling expressive, non-degenerate residual matrices while preserving training stability.
title Beyond the Birkhoff Polytope: Spectral-Sphere-Constrained Hyper-Connections
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2603.20896