On the Existence of Universal Lottery Tickets
This addresses the problem of efficient model reuse and initialization for practitioners in machine learning, though it is incremental as it builds on the lottery ticket hypothesis.
The paper theoretically proves the existence of universal lottery tickets—sparse subnetworks in randomly initialized deep neural networks that can be reused across tasks without further training, establishing their existence through formal proofs.
The lottery ticket hypothesis conjectures the existence of sparse subnetworks of large randomly initialized deep neural networks that can be successfully trained in isolation. Recent work has experimentally observed that some of these tickets can be practically reused across a variety of tasks, hinting at some form of universality. We formalize this concept and theoretically prove that not only do such universal tickets exist but they also do not require further training. Our proofs introduce a couple of technical innovations related to pruning for strong lottery tickets, including extensions of subset sum results and a strategy to leverage higher amounts of depth. Our explicit sparse constructions of universal function families might be of independent interest, as they highlight representational benefits induced by univariate convolutional architectures.