【深層学習】Attention - 全領域に応用され最高精度を叩き出す注意機構の仕組み【ディープラーニングの世界 vol. 24】#095 #VRアカデミア #DeepLearning

▼Topic This article explains RNNsearch, the ancestor of attention, a network that has seen explosive growth in Transformer and BERT. Attention not only achieved the phenomenal accuracy of GPT-3 in natural language processing, but is also being applied to a wide range of fields, including image processing and generative models. It's an essential element in any discussion of deep learning today! Check it out! ▼Related Playlist The World of Deep Learning    • Deep Learning の世界   Natural Language Processing Series    • 自然言語処理シリーズ   ▼Table of Contents (Added later. Please wait.) ▼References Bahdanau, Dzmitry, Kyunghyun Cho, and Yoshua Bengio. "Neural machine translation by jointly learning to align and translate." arXiv preprint arXiv:1409.0473 (2014). https://arxiv.org/abs/1409.0473 This is the original paper! It's written in a way that's not overly difficult to understand, including the history of the time, so it might be worth a read! [2019 Edition] Summary of Representative Natural Language Processing Models and Algorithms - Qiita https://qiita.com/LeftLetter/items/14... I use this as a reference for various videos. ▼Reference Videos RNN Video →    • 【深層学習】RNN の意味を徹底解説！【ディープラーニングの世界 vol. 8 】 ...   GRU Video →    • 【深層学習】GRU - RNN に記憶をもたせる試みその1【ディープラーニングの世界...   Three Ways to Use RNNs (For those who didn't understand the BiGRU part) →    • 【深層学習】RNN の3通りの使い方 - RNN の混乱ポイントを倒す！【ディープラ...   Bi-LSTM video (a companion to Bi-GRU) →    • 【深層学習】bi-LSTM - 前後の文脈を利用する Recurrent layer...   ▼In Closing Thank you for watching! If you enjoyed this video, please like and subscribe. If you have any questions or comments about the video, please leave them in the comments section or on Twitter! For business or collaboration requests, please send me a DM on Twitter. Video Creation: AIcia Solid (Twitter:   / aicia_solid  ) Video Editing: AIris Solid (Younger Sister) (Twitter:   / airis_solid  ) ======= Logo: TEICA (  / t_e_i_c_a  ) Model: http://3d.nicovideo.jp/works/td44519 Model by: W01fa (  / w01fa  )