安永會計師行今日三點半喺 YIA LT1 會有招聘講座, 現時任職 Senior III Auditor 嘅 #中大經濟 舊生 Samuel 亦都會在場同同學交流。有興趣全職或暑期實習嘅同學記得出席!
#Big4 #Audit #Intern #顯得我好矮
Did you know that one of the pioneers in neural network was an econometrician? This was not an easy book to find, but it was well worth it. The book is a collection of articles written by Halbert White—as in *White* Standard Errors—that rigorously laid the theoretical foundation for the effectiveness of neural nets, including universal approximation, consistency and asymptotic normality of learned parameters. These results fill a much-needed gap for those with a background in statistics.
deeplearning.ai
講開人工智能,業界泰斗吳恩達教授有關深度學習的免費網上課程在中國已經開始了,比美國早了近兩星期。若你有興趣認識人工智能為何物就萬不能錯過。
While we on the topic of AI, do you know that Andrew Ng’s online course on deep learning is already available in China, two weeks ahead of its official launch in the United States? It’s a very telling sign on how much a leader China is in the development of AI. I highly encourage anyone interested in knowing what AI really is about to take the course. You get to learn from one of the most important figures in the field, how much better than that can you get?
P.S. It’s also interesting that Andrew chose an economic topic as the very first example.
社評﹕AI革命席捲全球 政府須積極回應 – 20170901 – 社評 – 社評 – 明報新聞網
「人工智能」是今年最熱的議題,人云亦云的情況和三兩年前「大數據」不遑多讓。香港的「AI教育」其實不差:很多中小學都有Lego機械人隊。問題是學生無論多喜歡製作機械人,最終選擇以數據科學作為職業的數目相當少。這一來是因為香港的大學學位分配由政府主導,二來過去二十多年來政商界對科技的不重視,已經使香港的高科技產業遠遠落後週邊地區。
以香港科技積弱的程度,我相信新加坡和深圳一點都不擔心來自香港的競爭。當社會情況每況愈下,香港最大的優勢恐怕快要剩下作為走資港的地步了。
午夜凶鈴 THE RING
Japan’s Ball Robot on ISS
原來日本真係咗隻Haro上國際太空站,咁幾時有高達?
So #Haro actually exists on ISS, now give me my Gundam.
https://www.theverge.com/2017/7/17/15981250/japan-space-camera-drone-iss-int-ball
Windows Subsystem for Linux
Win10終於可以行bash,就可惜無得用GPU。
The best of both worlds. Unfortunately for data scientists (and cryptominers) there is no GPU support. #WindowsSubsystemForLinux #Ubuntu
MKL Limited to 1 Thread in Anaconda
If you have used the preload trick to use Math Kernel Library (MKL) with an early version of Anaconda, you will find that under Anaconda 4.4 MKL is limited a single thread no matter what settings you enter. The solution is to remove the LD_PRELOAD entry for MKL, log out, log back in and you should be good to go.