* The arXiv paper "Task Agnostic Continual Learning Using Online
Variational Bayes" is a preliminary pre-print of this paper. The main
differences between the versions are: 1. We develop new algorithmic framework
(FOO-VB). 2. We add multivariate Gaussian and matrix variate Gaussian
versions of the algorithm. 3. We demonstrate the new algorithm performance in
task agnostic scenarios Access Paper or Ask Questions
* This updates and significantly extends a previous article
(arXiv:1906.05827), Sections 6 and 7.1 are the most major additions. 30
pages. arXiv admin note: text overlap with arXiv:1906.05827 Access Paper or Ask Questions
* (1) A bug in the proof of implicit bias for matrix factorization was
fixed. v2 gives a characterization of the asymptotic bias of the factor
matrices, while v1 made a stronger claim on the limit direction of the
unfactored matrix. (2) v2 also includes new results on implicit bias of
mirror descent with realizable affine constraints Access Paper or Ask Questions
* Added empirical results of experiments on deep networks (Appendix E).
In addition, minor typos and phrasing mistakes were fixed Access Paper or Ask Questions
* Journal version (previous version appeared as conference paper in
ICLR ). Main improvements: We proved measure zero case for main theorem (with
implication for the rates), and the multi-class case. Both were not covered
in previous version Access Paper or Ask Questions