Skip To Main Content
Liu Ziyin on Muck Rack

Liu Ziyin

As seen in: Physical Review B
Is this you? As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work. Claim your profile

Get in touch with Liu

Contact Liu, search articles and posts on X, monitor coverage, and track replies from one place.

Learn more about Muck Rack

Actions

Share this page

Is this you?

As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.
Claim your profile

Articles

[2401.07085] Three Mechanisms of Feature Learning in a Linear Network

|
By Liu Ziyin
| arxiv.org
arXiv:2401.07085 (cs) [Submitted on 13 Jan 2024 (v1), last revised 21 Feb 2025 (this version, v3)] View PDF HTML (experimental) Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) Cite as: arXiv:2401.07085 [cs.LG] (or arXiv:2401.07085v3 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2401.07085 Submission history From: Yizhou Xu [ view email] [v1] Sat, 13 Jan 2024 14:21:46 UTC (1,259 KB) [v2] Sat, 4 May 2024 12:43:04 UTC (1,065 KB) [v3] Fri, 21 Feb 2025 11:50:09 UTC...
Open in Who Shared View Duplicates (56) Issue with article?

Zeroth, first, and second-order phase transitions in deep neural networks

|
By Liu Ziyin, Masahito Ueda
| Physical Review B
We investigate deep-learning-unique first-order and second-order phase transitions, whose phenomenology closely follows that in statistical physics. In particular, we prove that the competition between prediction error and model complexity in the training loss leads to the second-order phase transition for deep linear nets with one hidden layer and the first-order phase transition for nets with more than one hidden layer.
Open in Who Shared Issue with article?

Universal thermodynamic uncertainty relation in nonequilibrium dynamics

|
By Liu Ziyin, Masahito Ueda
| Physical Review B
Abstract We derive a universal thermodynamic uncertainty relation (TUR) that applies to an arbitrary observable in a general Markovian system.
Open in Who Shared Issue with article?
See all 4 articles