all AI news
Randomized K-FACs: Speeding up K-FAC with Randomized Numerical Linear Algebra. (arXiv:2206.15397v2 [cs.LG] UPDATED)
Sept. 29, 2022, 1:12 a.m. | Constantin Octavian Puiu
cs.LG updates on arXiv.org arxiv.org
K-FAC is a successful tractable implementation of Natural Gradient for Deep
Learning, which nevertheless suffers from the requirement to compute the
inverse of the Kronecker factors (through an eigen-decomposition). This can be
very time-consuming (or even prohibitive) when these factors are large. In this
paper, we theoretically show that, owing to the exponential-average
construction paradigm of the Kronecker factors that is typically used, their
eigen-spectrum must decay. We show numerically that in practice this decay is
very rapid, leading to …
More from arxiv.org / cs.LG updates on arXiv.org
Jobs in AI, ML, Big Data
Senior ML Researcher - 3D Geometry Processing | 3D Shape Generation | 3D Mesh Data
@ Promaton | Europe
Senior AI Engineer, EdTech (Remote)
@ Lightci | Toronto, Ontario
Data Scientist for Salesforce Applications
@ ManTech | 781G - Customer Site,San Antonio,TX
AI Research Scientist
@ Gridmatic | Cupertino, CA
Data Engineer
@ Global Atlantic Financial Group | Boston, Massachusetts, United States
Machine Learning Engineer - Conversation AI
@ DoorDash | Sunnyvale, CA; San Francisco, CA; Seattle, WA; Los Angeles, CA