3 problems

Matching

Marion et al.’s correlated-initialization conjecture for deep residual networks
For deep residual networks whose weights are initialized with correlations across layers, the infinite-depth limiting dynamics should interpolate continuously between the stochasti…
claimed progressdeep-learning-theory probability stochastic-analysis
03
Linear independence of polynomial compositions conjecture
For every fixed integer $m\ge 1$ and every collection of pairwise distinct nonconstant polynomials $f_1,\ldots,f_m\in\mathbb{C}[x]$ , there exists an integer $d_0$ such that, for ev…
claimed progressalgebra deep-learning-theory polynomial-identities
05
Progressive Feedforward Collapse conjecture for ResNet training
About 2 years old
Progressive Feedforward Collapse conjecture. At the terminal phase of ResNet training, neural collapse emerges in the last-layer features. Before the effective depth, there exists…
opendeep-learning-theory neural-collapse resnet
010

Marion et al.’s correlated-initialization conjecture for deep residual networks