How Do Training Methods Influence the Utilization of Vision Models?

TL;DR

This paper investigates how different training methods affect the importance of various layers in vision models, revealing that training regimes significantly influence layer criticality and utilization.

Contribution

It demonstrates that training procedures alter layer importance in vision models, providing new insights into neural network inner mechanics and training effects.

Findings

Improved training increases early layer importance.

Self-supervised training under-utilizes deeper layers.

Adversarial training shows opposite layer utilization trends.

Abstract

Not all learnable parameters (e.g., weights) contribute equally to a neural network's decision function. In fact, entire layers' parameters can sometimes be reset to random values with little to no impact on the model's decisions. We revisit earlier studies that examined how architecture and task complexity influence this phenomenon and ask: is this phenomenon also affected by how we train the model? We conducted experimental evaluations on a diverse set of ImageNet-1k classification models to explore this, keeping the architecture and training data constant but varying the training pipeline. Our findings reveal that the training method strongly influences which layers become critical to the decision function for a given task. For example, improved training regimes and self-supervised training increase the importance of early layers while significantly under-utilizing deeper layers. In contrast, methods such as adversarial training display an opposite trend. Our preliminary results extend previous findings, offering a more nuanced understanding of the inner mechanics of neural networks. Code: https://github.com/paulgavrikov/layer_criticality