Stories
Slash Boxes
Comments

SoylentNews is people

posted by Fnord666 on Wednesday August 15 2018, @05:44PM   Printer-friendly
from the faster-path-to-skynet dept.

Submitted via IRC for SoyCow1984

Students from Fast.ai, a small organization that runs free machine-learning courses online, just created an AI algorithm that outperforms code from Google's researchers, according to an important benchmark.

Fast.ai's success is important because it sometimes seems as if only those with huge resources can do advanced AI research.

Fast.ai consists of part-time students keen to try their hand at machine learning—and perhaps transition into a career in data science. It rents access to computers in Amazon's cloud.

But Fast.ai's team built an algorithm that beats Google's code, as measured using a benchmark called DAWNBench, from researchers at Stanford. This benchmark uses a common image classification task to track the speed of a deep-learning algorithm per dollar of compute power.

Google's researchers topped the previous rankings, in a category for training on several machines, using a custom-built collection its own chips designed specifically for machine learning. The Fast.ai team was able to produce something even faster, on roughly equivalent hardware.

"State-of-the-art results are not the exclusive domain of big companies," says Jeremy Howard, one of Fast.ai's founders and a prominent AI entrepreneur. Howard and his cofounder, Rachel Thomas, created Fast.ai to make AI more accessible and less exclusive.

Source: https://www.technologyreview.com/s/611858/small-team-of-ai-coders-beats-googles-code/


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 2) by MichaelDavidCrawford on Wednesday August 15 2018, @08:55PM

    by MichaelDavidCrawford (2339) Subscriber Badge <mdcrawford@gmail.com> on Wednesday August 15 2018, @08:55PM (#721915) Homepage Journal

    In many cases that works really well. For example my own creative commons piano EP [soggywizards.com]. We also have Vim and - dare I admit to it? - Emacs. There is GDB, GCC and the Linux kernel.

    But then we have gEdit as I experienced it the first time I tried to use it:

    It was only capable of holding the text that was visible in its window. I copied lots of text from some other program, pasted it into gEdit then quit that other app. When I scrolled down in my gEdit window the rest of that text was _gone_.

    Yet is was bundled with a popular Linux distro. I could name many other offenders.

    I have a good reason for my - usually but not always - preference for code that I pay for: if that code sucks I won't buy it anymore; neither will anyone else. Either its publisher fixes the problems or it goes out of business.

    I'm not saying I always prefer closed-source code - there is lots of Open Source and Free Software that I pay for, for example the copy of Vim that came with my MacBook Pro. I don't know about Apple's Vim in particular but it patches lots of OH HOW I HATE THIS WORD: FLOSS codebases before it ships then in macOS - the operating system formerly known by most as Mac Oh Ess Ecks but by Apple as Mac Oh Ess Ten.

    For the most part Apple does the right thing by distributing source [apple.com].

    --
    Yes I Have No Bananas. [gofundme.com]
    Starting Score:    1  point
    Karma-Bonus Modifier   +1  

    Total Score:   2