[NeurIPS 2018] How to Start Training: The Effect of Initialization and Architecture
This is a 3-minute video for the NeurIPS 2018 paper "How to Start Training: The Effect of Initialization and Architecture” by Boris Hanin and David Rolnick.
The conference presentation is at Poster #114, Wed, Dec. 5 @ 10:45-12:45.
Our full paper is available at:
https://arxiv.org/abs/1803.01719
Видео [NeurIPS 2018] How to Start Training: The Effect of Initialization and Architecture канала David Rolnick
The conference presentation is at Poster #114, Wed, Dec. 5 @ 10:45-12:45.
Our full paper is available at:
https://arxiv.org/abs/1803.01719
Видео [NeurIPS 2018] How to Start Training: The Effect of Initialization and Architecture канала David Rolnick
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
COMP-761: Lecture 5 (invariants and monovariants)COMP-761: Lecture 31 (dynamic programming)COMP-761: Lecture 14 (calculus II)COMP-761: Lecture 33 (computational complexity)COMP-761: Lecture 12 (graph theory II)COMP-761: Lecture 32 (miscellaneous)COMP-761: Lecture 38 (neural networks III)COMP-761: Lecture 13 (calculus I)COMP-761: Lecture 26 (graph algorithms III)COMP-761: Lecture 16 (Taylor series)COMP-761: Lecture 17 (probability I)COMP-761: Lecture 1 (intro)COMP-761: Lecture 28 (binary search trees I)COMP-761: Lecture 2 (proofs)COMP-761: Lecture 34 (computational complexity II)COMP-761: Lecture 6 (polynomials)COMP-761: Lecture 37 (neural networks II)COMP-761: Lecture 8 (combinatorics)COMP-761: Lecture 21 (linear programs II)COMP-761: Lecture 15 (calculus III)