Solving Large-scale Spatial Problems with Convolutional Neural Networks
This work addresses efficiency challenges in deep learning for large-scale spatial applications, representing an incremental improvement in transfer learning methods.
The paper tackles the computational intractability of large-scale spatial problems, specifically mobile infrastructure on demand (MID) with hundreds of agents, by using transfer learning with convolutional neural networks (CNNs) to improve training efficiency, achieving scalability with little to no performance degradation.
Over the past decade, deep learning research has been accelerated by increasingly powerful hardware, which facilitated rapid growth in the model complexity and the amount of data ingested. This is becoming unsustainable and therefore refocusing on efficiency is necessary. In this paper, we employ transfer learning to improve training efficiency for large-scale spatial problems. We propose that a convolutional neural network (CNN) can be trained on small windows of signals, but evaluated on arbitrarily large signals with little to no performance degradation, and provide a theoretical bound on the resulting generalization error. Our proof leverages shift-equivariance of CNNs, a property that is underexploited in transfer learning. The theoretical results are experimentally supported in the context of mobile infrastructure on demand (MID). The proposed approach is able to tackle MID at large scales with hundreds of agents, which was computationally intractable prior to this work.