Zobrazeno 1 - 1
of 1
pro vyhledávání: '"Kuru, N. G."'
Publikováno v:
Scopus-Elsevier
From CNNs to attention mechanisms, encoding inductive biases into neural networks has been a fruitful source of improvement in machine learning. Adding auxiliary losses to the main objective function is a general way of encoding biases that can help
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::af52613ad8885a2f6b6a0625c8a03e8a
