How to get good results with Deep Learning? Although there is no single straightforward way of tuning Deep Learning models, the experts from H2O.ai will share their practical experience in their talk from the Informatics evenings series. Not only will they explain the definition of Transfer Learning, but they will also discuss the most important hyperparameters one needs to tune and the order of tuning. Along the way, they will show a live demo in the no-code Deep Learning platform called H2O Hydrogen Torch. This lecture is conducted in English.
About the speakers
Yauhen Babakhin
Yauhen Babakhin currently works as a Principal Data Scientist at H2O.ai, where his focus area is Deep Learning and Computer Vision, in particular. He is a Kaggle competitions Grandmaster with a total of 12 gold medals in classic Machine Learning, Natural Language Processing and Computer Vision competitions. Currently, he has secured a position in the top-15 of the global Kaggle ranking.
Martin Barus
Martin Barus works as a Senior Data Scientist at H2O.ai alongside Yauhen. He graduated in Knowledge Engineering at FIT and is a Kaggle competitions Expert.
Adam Valenta
Adam Valenta is a Software Engineer at H2O.ai. Thanks to the faculty program of cooperation with industry, he started cooperating with this company in the last year of his studies at FIT to create his master thesis, Anomaly detection using Extended Isolation Forest, under the supervision of Veronika Maurerová also from H2O.ai.