When it comes to automating machine learning, Rachel Thomas notes, there are tradeoffs between the computationally...

When it comes to automating machine learning, Rachel Thomas notes, there are tradeoffs between the computationally…

Reading Time: 1 minute

When it comes to automating machine learning, Rachel Thomas notes, there are tradeoffs between the computationally intensive, “neural architecture search” and its alternative, “transfer learning.” Here’s a great overview of the tension between them.

http://www.fast.ai/2018/07/23/auto-ml-3/

31 comments

  1. MORAL is Momentary. Rest are only Hobbies

  2. this times are of great knowledge soon or later we will have that we seek and it will be know is it good or bad

  3. The way I see this, what Google announced was the development of a tool to partly automate and optimize an engineering task. Clearly this is the domain of companies: development of a new tool for their customers to use.

    This is clearly not research, i.e. work to create new knowledge, such as is performed at Stanford, Cambridge, Copenhagen University and Technical University of Denmark and many other places around the world.

    We should all keep this difference in mind, when we watch announcements from Google, Facebook, and others or attend their events – NOTE: Google I/O is not a conference with scientific presentation, but an event about new software developments by Google.

    PS: Would I like to attend an event arranged by Google? Clearly, yes. But I would NOT see that as participation in a scientific conference.

  4. Thank you for your sharing!

  5. Boy, if I was an editor for this for a general audience, I would require more examples. It’s almost unreadable to the general reader.

  6. Meg Tufano But this is not for a general audience, it is for the subgroup interested in artificial intelligence – however you define it.

  7. Niels Jensen If they want to solve the problems inherent in getting the public on their side, I’d suggest making sure they have a few paragraphs at the least for the general reader. It’s not that hard to do (I do it all day) (I’m a teacher ;’)).

  8. Meg Tufano But Google is not teaching they are selling to their customers, who are knowledgeable about AI and ML.

  9. Niels Jensen That’s what YOU wish . ;’) Trust me (and Gideon Rosenblatt will agree with me) that the people who make decisions about this stuff NEED help in understanding!

  10. Meg Tufano what is ML?

  11. ML machine Learning

  12. ‏الانضمام إلى المحادثة على Hangouts: hangouts.google.com – Google Hangouts

  13. 👍👍👍👍👍

  14. Why is it so difficult to implant any form of intelligence into people of power

    Why is it so difficult for people who work mob handed in gangs many to one

    Why is it so difficult to teach people in power to b humane

  15. bescuase they are bonded and held and on contract with their profession and they have it all already they are not going to take the chance to lose everything they worked hard on

  16. Eric Green thats really true….

  17. Sir I want job

  18. I have 10 Yeats experirince multinational company

  19. this is my goal plz help me how to do hacking plz tell me

  20. lol, I love automating machine learning, I’m in love with working that job :). I’m expert at machine learning, so that’s my job

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Sign up here for the latest articles. You can opt out at any time.


Subscribe by email:



Or subscribe by RSS:

Subscribe-by-RSS---Orange-Background
%d bloggers like this: