Diagnostic Visualization for Deep Neural Networks Using Stochastic Gradient Langevin Dynamics
This work addresses the challenge of diagnosing and debugging deep neural networks for researchers and practitioners, though it appears incremental as it builds on existing activation maximization methods.
The paper tackled the problem of interpreting deep neural networks by introducing Langevin Dynamics Activation Maximization (LDAM), a method that explores maximally activating pre-images and balances interpretability with pixel-level accuracy using a GAN-style discriminator, applied to MNIST, CIFAR, and ImageNet datasets.
The internal states of most deep neural networks are difficult to interpret, which makes diagnosis and debugging during training challenging. Activation maximization methods are widely used, but lead to multiple optima and are hard to interpret (appear noise-like) for complex neurons. Image-based methods use maximally-activating image regions which are easier to interpret, but do not provide pixel-level insight into why the neuron responds to them. In this work we introduce an MCMC method: Langevin Dynamics Activation Maximization (LDAM), which is designed for diagnostic visualization. LDAM provides two affordances in combination: the ability to explore the set of maximally activating pre-images, and the ability to trade-off interpretability and pixel-level accuracy using a GAN-style discriminator as a regularizer. We present case studies on MNIST, CIFAR and ImageNet datasets exploring these trade-offs. Finally we show that diagnostic visualization using LDAM leads to a novel insight into the parameter averaging method for deep net training.