no code implementations • 5 Jan 2021 • Kenya Tajima, Takahiko Henmi, Kohei Tsuchida, Esmeraldo Ronnie R. Zara, Tsuyoshi Kato
One of the two algorithms is based on the projected gradient method, in which each iteration of the projected gradient method takes $O(nd)$ computational cost and the sublinear convergence of the objective error is guaranteed.
no code implementations • 16 Aug 2020 • Takahiko Henmi, Esmeraldo Ronnie Rey Zara, Yoshihiro Hirohashi, Tsuyoshi Kato
In this study, we derived the initialization scheme again not from the simplified Kaiming model, but precisely from the modern CNN architectures, and empirically investigated how the new initialization method performs compared to the de facto standard ones that are widely used today.