On the Existence of Universal Lottery Tickets

Burkholz, Rebekka and Laha, Nilanjana and Mukherjee, Rajarshi and Gotovos, Alkis
(2022) On the Existence of Universal Lottery Tickets.
In: The Tenth International Conference on Learning Representations, 25-29 Apr 2022, virtual.
Conference: ICLR International Conference on Learning Representations


Download (976kB) | Preview
Official URL: https://openreview.net/forum?id=SYB4WrJql1n


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.


Actions (login required)

View Item View Item