← Search

Chenggang Zhao

2 accepted papers

2026

mHC: Manifold-Constrained Hyper-Connections

ICML 2026spotlight

Recently, studies exemplified by Hyper-Connections (HC) have extended the ubiquitous residual connection paradigm established over the past decade by expanding the residual stream width and diversifying connectivity patterns. While yielding substantial performance gains, this diversification fundame…

Cited by 0SourceScholar
2024

DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models

ACL 2024long

In the era of large language models, Mixture-of-Experts (MoE) is a promising architecture for managing computational costs when scaling up model parameters. However, conventional MoE architectures like GShard, which activate the top-K out of N experts, face challenges in ensuring expert specializati…