We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Linear Mode Connectivity and the Lottery Ticket Hypothesis

Abstract: We introduce "instability analysis," which assesses whether a neural network optimizes to the same, linearly connected minimum under different samples of SGD noise. We find that standard vision models become "stable" in this way early in training. From then on, the outcome of optimization is determined to within a linearly connected region. We use instability to study "iterative magnitude pruning" (IMP), the procedure used by work on the lottery ticket hypothesis to identify subnetworks that could have trained to full accuracy from initialization. We find that these subnetworks only reach full accuracy when they are stable, which either occurs at initialization for small-scale settings (MNIST) or early in training for large-scale settings (Resnet-50 and Inception-v3 on ImageNet).
This submission subsumes 1903.01611 ("Stabilizing the Lottery Ticket Hypothesis" and "The Lottery Ticket Hypothesis at Scale")
Comments: This submission subsumes 1903.01611 ("Stabilizing the Lottery Ticket Hypothesis" and "The Lottery Ticket Hypothesis at Scale")
Subjects: Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE); Machine Learning (stat.ML)
Cite as: arXiv:1912.05671 [cs.LG]
  (or arXiv:1912.05671v2 [cs.LG] for this version)

Submission history

From: Jonathan Frankle [view email]
[v1] Wed, 11 Dec 2019 22:22:21 GMT (1980kb,D)
[v2] Thu, 20 Feb 2020 20:39:29 GMT (1780kb,D)

Link back to: arXiv, form interface, contact.