Visualizing the PHATE of Neural Networks

Abstract

Understanding why and how certain neural networks outperform others is key to guiding future development of network architectures and optimization methods. To this end, we introduce a novel visualization algorithm that reveals the internal geometry of such networks: Multislice PHATE (M-PHATE), the first method designed explicitly to visualize how a neural network's hidden representations of data evolve throughout the course of training. We demonstrate that our visualization provides intuitive, detailed summaries of the learning dynamics beyond simple global measures (i.e., validation loss and accuracy), without the need to access validation data. Furthermore, M-PHATE better captures both the dynamics and community structure of the hidden units as compared to visualization based on standard dimensionality reduction methods (e.g., ISOMAP, t-SNE). We demonstrate M-PHATE with two vignettes: continual learning and generalization. In the former, the M-PHATE visualizations display the mechanism of "catastrophic forgetting" which is a major challenge for learning in task-switching contexts. In the latter, our visualizations reveal how increased heterogeneity among hidden units correlates with improved generalization performance. An implementation of M-PHATE, along with scripts to reproduce the figures in this paper, is available at https://github.com/scottgigante/M-PHATE.

References in corpus (8)

• Scikit-learn: Machine Learning in Python
• On the Number of Linear Regions of Deep Neural Networks
• A Closer Look at Memorization in Deep Networks
• Re-evaluating Continual Learning Scenarios: A Categorization and Case for Strong Baselines
• Qualitatively characterizing neural network optimization problems
• Learning and Generalization in Overparameterized Neural Networks, Going Beyond Two Layers
• Understanding the geometry of transport: diffusion maps for Lagrangian trajectory data unravel coherent sets
• Hierarchical Coupled Geometry Analysis for Neuronal Structure and Activity Pattern Discovery

Cited by in corpus (3)

• Neural population geometry: An approach for understanding biological and artificial neural networks
• Extendable and invertible manifold learning with geometry regularized autoencoders
• Dropout as a Regularizer of Interaction Effects