【新智元导读】2026年架构革命的枪声已经打响！ResNet用十年证明了「加法捷径」能救训练，但也暴露了「只加不减」的天花板。DeepSeek新年王炸之后，普林斯顿和UCLA新作DDL让网络学会忘记、重写和反转。

作为计算机视觉领域的顶尖科学家，何恺明毕业于清华大学物理系，现任麻省理工学院电气工程与计算机科学系副教授（终身教职）。其最著名的成就，就是提出了ResNet架构,解决了深度神经网络中“越深越难训练”的根本性问题，ResNet也被称为现代深度学习发展史 ...

DeepSeek has published a technical paper co-authored by founder Liang Wenfeng proposing a rethink of its core deep learning ...

DeepSeek’s proposed ‘mHC’ design could change how AI models are trained, but experts caution it still needs to prove itself ...

Understand how 1x1 convolutions work and why they’re essential in modern neural network architectures like ResNet and ...