Understanding intermediate layers using linear classifier probes

Guillaume Alain, Yoshua Bengio

TL;DR: Investigating deep learning models by proposing a different concept of information

Abstract:

Neural network models have a reputation for being black boxes.
We propose a new method to better understand the roles and dynamics
of the intermediate layers.

Our method uses linear classifiers, referred to as "probes",
where a probe can only use the hidden units of a given intermediate layer
as discriminating features.
Moreover, these probes cannot affect the training phase of a model,
and they are generally added after training.

We demonstrate how this can be used to develop a better intuition
about models and to diagnose potential problems.

Keywords: Deep learning, Supervised Learning, Theory

Conflicts: umontreal.ca, google.com