paper-with-me

홈 › Papers

Layer rotation: a surprisingly simple indicator of generalization in deep networks?

2019-05-17 · ICML Workshop Deep_Phenomen 2019 6 · Simon Carbonnelle, Christophe De Vleeschouwer

Our work presents empirical evidence that layer rotation, i.e. the evolution across training of the cosine distance between each layer's weight vector and its initialization, constitutes an impressively consistent indicator of generalization performance. Compared to previously studied indicators of generalization, we show that layer rotation has the additional benefit of being easily monitored and controlled, as well as having a network-independent optimum: the training procedures during which all layers' weights reach a cosine distance of 1 from their initialization consistently outperform other configurations -by up to 20% test accuracy. Finally, our results also suggest that the study of layer rotation can provide a unified framework to explain the impact of weight decay and adaptive gradient methods on generalization.

📄 PDF Abstract BibTeX

Code (1)

anonymousdlresearch/layer-rotation 공식 구현 tf

Similar Papers 제목 키워드 기반

Layer rotation: a surprisingly powerful indicator of generalization in deep networks?

2018-06-05 · Simon Carbonnelle, Christophe De Vleeschouwer

Our work presents extensive empirical evidence that layer rotation, i.e. the evolution across training of the cosine distance between each layer's weight vector and its initialization, constitutes an impressively consist…

Generalization to Out-of-Distribution transformations

2021-09-29 · Shanka Subhra Mondal, Zack Dulberg, Jonathan Cohen

Humans understand a set of canonical geometric transformations (such as translation, rotation and scaling) that support generalization by being untethered to any specific object. We explored inductive biases that allowed…

DiversityTranslation

An experimental study of layer-level training speed and its impact on generalization

2018-09-27 · Simon Carbonnelle, Christophe De Vleeschouwer

How optimization influences the generalization ability of a DNN is still an active area of research. This work aims to unveil and study a factor of influence: the speed at which each layer trains. In our preliminary work…

Who Handles Orientation? Investigating Invariance in Feature Matching

2026-04-13 · David Nordström, Johan Edstedt, Fredrik Kahl, Georg Bökman arxiv

Finding matching keypoints between images is a core problem in 3D computer vision. However, modern matchers struggle with large in-plane rotations. A straightforward mitigation is to learn rotation invariance via data au…

Data AugmentationImage Matching

A case where a spindly two-layer linear network whips any neural network with a fully connected input layer

2020-10-16 · Manfred K. Warmuth, Wojciech Kotłowski, Ehsan Amid

It was conjectured that any neural network of any structure and arbitrary differentiable transfer functions at the nodes cannot learn the following problem sample efficiently when trained with gradient descent: The insta…