[Submitted on 9 Jan 2025]
View a PDF of the paper titled Emergent weight morphologies in deep neural networks, by Pascal de Jong and 2 other authors
Abstract:Whether deep neural networks can exhibit emergent behaviour is not only relevant for understanding how deep learning works, it is also pivotal for estimating potential security risks of increasingly capable artificial intelligence systems. Here, we show that training deep neural networks gives rise to emergent weight morphologies independent of the training data. Specifically, in analogy to condensed matter physics, we derive a theory that predict that the homogeneous state of deep neural networks is unstable in a way that leads to the emergence of periodic channel structures. We verified these structures by performing numerical experiments on a variety of data sets. Our work demonstrates emergence in the training of deep neural networks, which impacts the achievable performance of deep neural networks.
Submission history
From: Steffen Rulands Prof. Dr. [view email]
[v1]
Thu, 9 Jan 2025 19:48:51 UTC (1,509 KB)
Source link
lol
