Тёмный
No video :(

Yann LeCun | May 18, 2021 | The Energy-Based Learning Model 

Mathematical Picture Language
Подписаться 1,8 тыс.
Просмотров 18 тыс.
50% 1

Опубликовано:

 

28 авг 2024

Поделиться:

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист
Посмотреть позже
Комментарии : 8   
@maxwellsdaemon7
@maxwellsdaemon7 3 года назад
At 8:17, the indices on the second and third lines look incorrect, shouldn't they be g_k, z_k, w_k instead of g_{k-1}, z_{k-1}, w_{k-1} in g?
@ramuk1127
@ramuk1127 2 года назад
No, look at the first line. Since z_k+1 is equivalent to g_k(z_k, w_k) it has to follow that the jacobian in the 2nd and 3rd lines can be written to compensate for backpropagation
@maxwellsdaemon7
@maxwellsdaemon7 2 года назад
@@ramuk1127 take a look at the previous slide and take the gradient with respect to z_k.
@mathematicalpicturelanguag2715
@mathematicalpicturelanguag2715 3 года назад
For more videos from the Mathematical Picture Language Tuesday seminar, see ru-vid.com/show-UCrlS3CuPlahBp_M46fDaWVwvideos
@carlotonydaristotile7420
@carlotonydaristotile7420 2 года назад
Tiny insert brains, like a jumping spider has a 3d model of the world.
@MichaelBryantthefirstangel
@MichaelBryantthefirstangel 3 года назад
Energy VS Agency VS Time. If Agency could only be quantified…
Далее
Can You Bend This Bar?
01:00
Просмотров 4,7 млн
GEOMETRIC DEEP LEARNING BLUEPRINT
3:33:23
Просмотров 178 тыс.
This is why Deep Learning is really weird.
2:06:38
Просмотров 384 тыс.
Why Does Diffusion Work Better than Auto-Regression?
20:18
Diffusion and Score-Based Generative Models
1:32:01
Просмотров 74 тыс.
Can You Bend This Bar?
01:00
Просмотров 4,7 млн