
許多人寫了深度學(xué)習(xí)將如何掌控世界的文章,以及為什么要重視深度學(xué)習(xí);我無法一一回應(yīng)他們。 使用深度學(xué)習(xí)是最讓我覺得變成了一個魔術(shù)師的事。最近我清楚地看到,每一個偉大的軟件將在未來3年內(nèi)由深度學(xué)習(xí)驅(qū)動發(fā)展。 然而,當下深度學(xué)習(xí)仍不是最主流,所以我想分享一些非常有才華的貢獻者的工作,希望能夠使深度學(xué)習(xí)離大家更接近一點。
Here’s ten reasons why I think deep learning is living up to the hype…

…it’s equally impressive for anime art

2 模仿人類語音閱讀 Deep learning has enabled a text-to-speech system that is almostindistinguishable from human voice. Think of the possibilities!
https://deepmind.com/blog/wavenet-generative-model-raw-audio/
https://soundcloud.com/oliver-cameron-330659812/wavenet-sample-2

3 創(chuàng)作經(jīng)典音樂 Deep learning can compose classical music that you’d believe to be created by a human
https://soundcloud.com/oliver-cameron-330659812/wavenet-sample-1

4 模仿畫家的畫風(fēng)作畫 Deep learning can replicate the style of your favorite painter with the image of your choice

這種DL技術(shù)的作畫論文在這里.
5 使用深度學(xué)習(xí)創(chuàng)作字體 Deep learning can even generate fonts!

6 深度學(xué)習(xí)修補圖片缺失部分 Deep learning can auto-fill missing parts of an image, by predicting what should be in that space

7 讓機器人訓(xùn)練成直立行走 Deep learning can train a robot to walk like a human

或者像人一樣拿東西…or train a robot to grasp objects like a human
8 為圖片自動取一個標題 Deep learning can caption an image, just like a human would

9 想要畫一幅漂亮的風(fēng)景畫? Want to sketch a beautiful landscape but can’t draw? Don’t worry, deep learning can take it from here!

10 讓深度學(xué)習(xí)來開車! Best of all (at least to me), you can train a deep neural network to steer a car**, just like a human

為什么我確信深度學(xué)習(xí)是革命性的? 如果你仔細看看機器正在執(zhí)行的上述任務(wù),你會看到一個共同的特性:創(chuàng)造力。 機器,雖然通過人類給予的數(shù)據(jù)進行訓(xùn)練,能夠展示接近真實藝術(shù)的性格。
我們花了幾十年的時間建設(shè)和優(yōu)化機器,讓其計算能力更為強大。 在接下來的十年中,我希望我們能夠發(fā)現(xiàn),當我們的軟件創(chuàng)造性地思考時,人類將能夠被賦予什么能力。
Ref
https://medium.com/@olivercameron/deep-learning-is-revolutionary-d0f3667bafa0#.av4uxfuwj