Critical Percolation as a Framework to Analyze the Training of Deep Networks
arXiv:1802.02154
Abstract
In this paper we approach two relevant deep learning topics: i) tackling of graph structured input data and ii) a better understanding and analysis of deep networks and related learning algorithms. With this in mind we focus on the topological classification of reachability in a particular subset of planar graphs (Mazes). Doing so, we are able to model the topology of data while staying in Euclidean space, thus allowing its processing with standard CNN architectures. We suggest a suitable architecture for this problem and show that it can express a perfect solution to the classification task. The shape of the cost function around this solution is also derived and, remarkably, does not depend on the size of the maze in the large maze limit. Responsible for this behavior are rare events in the dataset which strongly regulate the shape of the cost function near this global minimum. We further identify an obstacle to learning in the form of poorly performing local minima in which the network chooses to ignore some of the inputs. We further support our claims with training experiments and numerical analysis of the cost function on networks with up to layers.
arXiv commentAccepted to ICLR 2018 as a conference paper
References in corpus (13)
- Geometric deep learning: going beyond Euclidean data
- Spectral Networks and Locally Connected Networks on Graphs
- Deep Convolutional Networks on Graph-Structured Data
- Understanding deep learning requires rethinking generalization
- Identifying and attacking the saddle point problem in high-dimensional non-convex optimization
- Gated Graph Sequence Neural Networks
- All you need is a good init
- Toward Deeper Understanding of Neural Networks: The Power of Initialization and a Dual View on Expressivity
- The Predictron: End-To-End Learning and Planning
- Shortest-Path Fractal Dimension for Percolation in Two and Three Dimensions
- Value Prediction Network
- Local minima in training of neural networks
- Singular Behavior of Eigenstates in Anderson's Model of Localization