DEEP LEARNING AND ITS BIGGEST ACHIEVEMENTS

Seria: 
Seminarium Zakładu Fizyki Teoretycznej
Prelegent i afiliacja: 
mgr Szymon NAKONECZNY, Narodowe Centrum Badań Jądrowych
Data: 
śr., 2018-11-21 12:15
Miejsce: 
pawilon NCBJ, sala 22, Hoża 69
Streszczenie: 

The first neural network was invented back in 1957, however, it is the last years when machine learning models achieved the most important breakthrough with the development of deep learning. This led to creation of AI algorithms which defeated humans in popular and complicated games like chess and Go, allowed us to communicate with digital devices through our own natural language and made advertisement companies earn tons of money through personalized ads. During the seminar, I will introduce the key concepts behind the deep learning models. This will include basics of building a neural network, the extension to deep convolutional models, as well as word embedding algorithms which are now one of the first steps of Natural Language Processing. Finally, I will describe various deep network architectures and their practical applications.

                                                                                             Serdecznie zapraszamy,
M. Kowal, W. Piechocki, J. Skalski, L. Szymanowski
Załączniki: