Microsoft Research ブログ
読み込み中…
Microsoft Research ブログ
Microsoft and NVIDIA introduce parameter-efficient multimodal transformers for video representation learning
| Yale Song
Understanding video is one of the most c…
Microsoft Research ブログ
Conversations with data: Advancing the state of the art in language-driven data exploration
| Alex Polozov, Chris Meek, と Ahmed Awadallah
One key aspiration of AI is to develop n…
Microsoft Research ブログ
Factorized layers revisited: Compressing deep networks without playing the lottery
| Misha Khodak, Neil Tenenholtz, Lester Mackey, と Nicolo Fusi
From BiT (928 million parameters (opens …