Videolibrary
RUS  ENG    JOURNALS   PEOPLE   ORGANISATIONS   CONFERENCES   SEMINARS   VIDEO LIBRARY   PACKAGE AMSBIB  
Video Library
Archive
Most viewed videos

Search
RSS
New in collection






Applied Mathematics Day
September 22, 2017 17:30–17:40, Moscow, 8 Gubina
 


Deep learning may need new mathematics

Xuecang Zhang

Huawei Technologies Co., Ltd., Seoul
Video records:
MP4 372.1 Mb
MP4 101.9 Mb

Number of views:
This page:359
Video files:105

Xuecang Zhang



Abstract: Many questions related to deep learning can be formulated and analyzed mathematically, but the theoretical understanding of deep learning is very limited. For example, there is definitely no mathematical theory that explains why deep learning works so well. Sometimes ago, the neural networks topic is abandoned by mathematicians. It as a research desert and is not mathematically beautiful. But in current AI wave, deep neural networks work very well in many domains, maybe we need to reconsider deep learning from mathematics point of view.

Language: English
 
  Contact us:
 Terms of Use  Registration to the website  Logotypes © Steklov Mathematical Institute RAS, 2024