View PDF ↗
PDF Viewer

Loading PDF...

This may take a moment

Open Full PDF

BUILDER'S SANDBOX

Core Pattern

AI-generated implementation pattern based on this paper's core methodology.

Understand the commercial significance and market impact.

7-day free trial. Cancel anytime.

Estimated $10K - $14K over 6-10 weeks.

See exactly what it costs to build this -- with 3 comparable funded startups.

7-day free trial. Cancel anytime.

Discover the researchers behind this paper and find similar experts.

7-day free trial. Cancel anytime.

Founder's Pitch

"Develop a stability-enhancing framework for deep learning architectures using orthogonal hyper-connections."

LLM TrainingScore: 4View PDF ↗

Commercial Viability Breakdown

0-10 scale

High Potential

1/4 signals

2.5

Quick Build

2/4 signals

5

Series A Potential

0/4 signals

0

Explore the full citation network and related research.

7-day free trial. Cancel anytime.

Understand the commercial significance and market impact.

7-day free trial. Cancel anytime.

Get detailed profiles of the research team.

7-day free trial. Cancel anytime.

References (32)

[1]
KromHC: Manifold-Constrained Hyper-Connections with Kronecker-Product Residual Matrices
2026Wuyang Zhou, Yuxuan Gu et al.
[2]
mHC: Manifold-Constrained Hyper-Connections
2025Zhenda Xie, Yixuan Wei et al.
[3]
Hierarchical Reasoning Model
2025Guan Wang, Jin Li et al.
[4]
ARC-AGI-2: A New Challenge for Frontier AI Reasoning Systems
2025Francois Chollet, Mike Knoop et al.
[5]
Hyper-Connections
2024Defa Zhu, Hongzhi Huang et al.
[6]
Noise Is Not the Main Factor Behind the Gap Between SGD and Adam on Transformers, but Sign Descent Might Be
2023Frederik Kunstner, Jacques Chen et al.
[7]
Infeasible Deterministic, Stochastic, and Variance-Reduction Algorithms for Optimization under Orthogonality Constraints
2023Pierre Ablin, Simon Vary et al.
[8]
A Unified Framework for Implicit Sinkhorn Differentiation
2022Marvin Eisenberger, Aysim Toker et al.
[9]
DeepNet: Scaling Transformers to 1,000 Layers
2022Hongyu Wang, Shuming Ma et al.
[10]
Tensor Programs III: Neural Matrix Laws
2020Greg Yang
[11]
Efficient Riemannian Optimization on the Stiefel Manifold via the Cayley Transform
2020Jun Li, Fuxin Li et al.
[12]
Trivializations for Gradient-Based Optimization on Manifolds
2019Mario Lezcano Casado
[13]
Operations on certain non-commutative operator-valued random variables
2018D. Voiculescu, D. Voiculescu
[14]
Dynamical Isometry is Achieved in Residual Networks in a Universal Way for any Activation Function
2018W. Tarnowski, P. Warchol et al.
[15]
Visualizing the Loss Landscape of Neural Nets
2017Hao Li, Zheng Xu et al.
[16]
Mean Field Residual Networks: On the Edge of Chaos
2017Greg Yang, S. Schoenholz
[17]
Resurrecting the sigmoid in deep learning through dynamical isometry: theory and practice
2017Jeffrey Pennington, S. Schoenholz et al.
[18]
Attention is All you Need
2017Ashish Vaswani, Noam Shazeer et al.
[19]
Deep Information Propagation
2016S. Schoenholz, J. Gilmer et al.
[20]
Full-Capacity Unitary Recurrent Neural Networks
2016Scott Wisdom, Thomas Powers et al.

Showing 20 of 32 references