Deep learning of the neural networks has become a breakthrough in the AI systems advance. However, some obstacles limit their practical application. At AI Conference speaker Dmitry Babaev will reveal the key problems of deep neural networks and possible solutions to them.
The speaker is a researcher at Sberbank AI laboratory. His main task lies in application of new machine learning methods in terms of classic business problems.
Previously, the speaker worked in such companies as МТS, Tinkoff Bank and Yandex:
- at MTS, he was Data Science Lead and developed this direction in the company;
- at Yandex, he developed Suggest, the first version of the autocomplete for user’s searches.
At the conference, Dmitry Babaev will provide the detailed information about such problems of the today’s neural networks generation as a large data volume of marked data necessary for network learning as well as difficulties of experience sharing of the trained network among other tasks.