Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
×
Microsoft Businesses Cloud IT

How Machine Learning Ate Microsoft 96

snydeq writes Yesterday's announcement of Azure Machine Learning offers the latest sign of Microsoft's deep machine learning expertise — now available to developers everywhere, InfoWorld reports. "Machine learning has infiltrated Microsoft products from Bing to Office to Windows 8 to Xbox games. Its flashiest vehicle may be the futuristic Skype Translator, which handles two-way voice conversations in different languages. Now, with machine learning available on the Azure cloud, developers can build learning capabilities into their own applications: recommendations, sentiment analysis, fraud detection, fault prediction, and more. The idea of the new Azure offering is to democratize machine learning, so you no longer need to hire someone with a doctorate to use a machine learning algorithm."
This discussion has been archived. No new comments can be posted.

How Machine Learning Ate Microsoft

Comments Filter:
  • by Anonymous Coward on Friday February 20, 2015 @05:35AM (#49093475)

    Ate or Aid?

  • by Anonymous Coward
    ... this is just their latest way to get their hands on your data.
    • by Anonymous Coward

      Well, of course. Businesses don't do anything for free, so you're not paranoid, just cautious.

      • by Anonymous Coward

        Don't be silly, Microsoft are in the business of selling server licenses. Many many businesses are asking for an on-premises versions of the azure tools, and this is just the publicly available "trial version".

        • by Lennie ( 16154 )

          I doubt it. They are in the business of selling products and services, they don't care what they can sell. They are a business trying to make money and stay relevant.

          If running a porn streaming service wouldn't damage their image and was something they thought they knew how to run well and make good money on, I'm sure they would just add it to their list of services.

          Now to be a bit more specific, of course they want your data. You see this happening especially on the consumer side.

          For example: where can I g

          • by jbolden ( 176878 ) on Friday February 20, 2015 @08:39AM (#49093959) Homepage

            For example: where can I get a copy of SkyDrive/OneDrive/whatever which I can run on my own systems ?

            SharePoint will do that.
            For that matter you can run the entire Azure suite in your private location: http://www.microsoft.com/en-us... [microsoft.com]

            • by Lennie ( 16154 )

              But important parts are missing.

              Some examples:
              - AzureAD, specifically ACS
              - Site Recovery for disaster recovery

              These are all online services with no buy/download equivalent from Microsoft.

              • by jbolden ( 176878 )

                Originally we were talking about some fairly mundane features. You aren't. You are talking about stuff that really is fairly complex, I'd say more SaaS than IaaS/PaaS.

                Site Recovery is an Azure public cloud recovery solution. If I'm running a private copy of Azure then either:

                a) I want to recover to the public Azure in which case Microsoft's Site Recovery works fine
                b) I want to recover to another private data center in which I want to use a clustering / replication strategy.

                AzureAD requires contracts a

          • by Lennie ( 16154 ) on Friday February 20, 2015 @08:42AM (#49093975)

            Let me be clear: what applies to Azure as a foreigner applies also to Amazon/AWS, Google, Rackspace, IBM/SoftLayer, CenturyLink, DigitalOcean, Vultr, Linode, PeerOne or any other US-based company (even if they run the service in Europe for example).

            But I noticed there are others in the world, for example on the OpenStack Marketplace:
            http://www.openstack.org/marke... [openstack.org]

    • ... this is just their latest way to get their hands on your data.

      Unfortunately, they're all like this. Pretty soon, Facebook is gonna want to know who my friends are. After they get that data, they'll be using the data I report about what I had for breakfast to show my friends ads. Then, they'll be providing "Like" buttons to report the following breakfast data from my friends back to me: "Hey Mikey, he likes it!"

    • Just because you're paranoid doesn't mean they aren't out to get you.
  • How Bing learns (Score:3, Interesting)

    by lu-darp ( 469705 ) on Friday February 20, 2015 @05:54AM (#49093537)

    So how did MS apply "machine learning" to make Bing not suck? By holding an internal competition to see who's algorithm processed "user improvement program" data best. So that essentially meant training it up to match Google search results (and presumably, which links "consenting" users clicked).

    (OK, I'm sure they've come a long way since then on their own merits, but we can't let them live that one down ;-)

    • Re:How Bing learns (Score:5, Insightful)

      by brunes69 ( 86786 ) <slashdot@[ ]rstead.org ['kei' in gap]> on Friday February 20, 2015 @07:38AM (#49093749)

      The problem with your statement is you make the assertion that Bing no longer sucks, which is false. Bing is still horrible.

      • I dunno...I just put the following query into Bing: "is bing horrible", and it came up with Why Bing Sucks. Top 5 Reasons [geeknizer.com]. So...it sucks, but it certainly isn't horrible.

        • Re:How Bing learns (Score:5, Interesting)

          by Marginal Coward ( 3557951 ) on Friday February 20, 2015 @09:19AM (#49094155)

          Although the above was just a joke, I actually clicked on the link after I submitted it, and it turns out to be an old page from 2009. It provides the follows searches which it says "just don't work" on Bing (in 2009):

          “Was Einstein married?”
          “What did Benjamin Franklin invent?”
          “What is the top selling album of all time?”

          I did a quick comparison of those three between Bing and Google, and the results seemed pretty comparable. In fact, I thought Bing did a little better on the first two, and Google did a little better on the last one - primarily because it provided a nice blurb from Wikipedia in the results.

          So, although I think we can all agree that Bing was "horrible" in the past, it's come a long ways. It's not like in the old days when Google was clearly the best - I think you could use any of the major search engines now and do just fine.

          • Re:How Bing learns (Score:4, Interesting)

            by lgw ( 121541 ) on Friday February 20, 2015 @10:54AM (#49094571) Journal

            DuckDuckGo uses Bing, but it does some nice enhancements, like putting Wikipedia results first very often. If you want to live a Google-free life, and don't want to give your search history to MS either - ddg.gg [duckduckgo.com].

            • by doom ( 14564 )
              Also: blekko.com, startpage.com. There's no particular need to use google for web searches any more.
              • by lgw ( 121541 )

                Startpage is google, right? Sadly, it's blocked at work for me, but I don't even like to give traffic to google - who knows what they can figure out.

            • Comment removed based on user account deletion
              • Ah, but DDG has !bangs [duckduckgo.com], so you can... duck?... for "!g foo" to get the Google results instead. I spent a few days acclimating to DDG and now use it for almost everything, falling back to Google for the 1% of the time when I don't get the results I expect. Also works for a few hundred other things, including the old green mare herself: "!/. foo" searches Slashdot.

              • by lgw ( 121541 )

                I've never noticed the difference in search results. Maybe I just don't care. Or maybe Google's "search bubble", where they give you results based on creepily stalking your search history works out well for you and prevents you from seeing results you don't like.

                I missed the calculator built into Google at first, but then I found that with ddg you can just do !wa for wolfram alpha, which is more comprehensive than Google calculator. "!wa speed of light in furlongs per fortnight" works as well as Google,

          • So, although I think we can all agree that Bing was "horrible" in the past, it's come a long ways.

            I don't know man, every once in a while I use Bing to search for stuff, and it seems pretty generally bad. Maybe someone at Microsoft read that page and optimised those special cases.

    • Did Bing suck. I did the Bing vs. Google head to head test about 18 month back a few times. (looks like it might still be online at: http://www.bingiton.com/ [bingiton.com] And I most typically scored 3-Google, 2-Bing with often Bing having some interesting results Google didn't have. For example Bing tends to do better in hitting a better diversity of current information. Bing may be a bit behind possibly and I'm not even comfortable saying that, but sucks no.

      • Re: (Score:2, Interesting)

        by Anonymous Coward
        To me it seems more like Google has got _worse_ than Bing has got better.

        Too often when I put in very specific keywords that very relevant to a particular article Google doesn't show the most likely hit, it instead tries to suggest what Joe Six pack might want and that I actually made typos to the keywords or typed those keywords by mistake and ignores some of them.

        Then when I turn on Verbatim on Google, it gives me a list of all the pages that contain those keywords in a useless order (which could include
  • Dammit! (Score:5, Funny)

    by nospam007 ( 722110 ) * on Friday February 20, 2015 @06:07AM (#49093571)

    "recommendations, sentiment analysis, fraud detection, fault prediction, and more."

    I'll never get a job again if they use that in interviews.

    • by Anonymous Coward

      *Must have 10 years experience in Azure Machine Learning.

  • by ProzakLord ( 1087161 ) on Friday February 20, 2015 @06:10AM (#49093581)
    Baing about to finish a PhD this is worrying thinking that the deep understanding of a technique can be replaced by a well programmed API. But of course managerial people, a.k.a decission makers will eat that raw. If you want my data just ask for it.
    • If you want my data just ask for it.

      OK, let's start with this: is "ProzakLord" your real first name or your last? Either way, can you provide us with the other one?

    • Sure. With the new API they will know what kernel to apply to their SVM learner, and the system will set the correct data types and preprocess the data just in the way that a parametric model needs, and will know when to segment your data when it is appropriate to apply different models to each segment, and will consider misclassification costs, and will know how to handle unbalanced datasets, and how to prevent overfitting, and when it is ok to apply ensemble models, and will write a summary justifying the
      • Now, now... Don't be petty. I'm sure this will let all the BI "analysts" make the same mistakes they always do, but much more cheaply.

        The flaw in your knowledge behind your well put, albeit sarcastic, commentary is the fact that most people don't care about quality and actively dislike information that rebuts their previously-held conclusions (like well-analyzed data is wont to do). As such, quality data analysis is greatly overvalued by folks who do it, while undervalued or misused by those who most need i

    • If you want my data just ask for it.

      What data?

  • by louic ( 1841824 ) on Friday February 20, 2015 @06:16AM (#49093595)
    I foresee a bright future where lots of correlations will be found. Without "a PhD" or someone who knows what they are doing ALL software is worthless.
    • Now at least we have a tool to tell them apart.

      http://science.slashdot.org/st... [slashdot.org]

    • by gruntkowski ( 1743014 ) on Friday February 20, 2015 @06:47AM (#49093671)
      That is so correct. I've tested the machine learning software on Azure. It is very nice and quite powerful. But without knowing what you are really doing, you probably get results which seem nice, but are in fact complete bogus. If you do not know what overfitting is for example, good luck using machine learning algorithms. If some manager starts using this, may god have mercy on us all...
      • by Dr_Barnowl ( 709838 ) on Friday February 20, 2015 @07:36AM (#49093745)

        Sounds like Reason [bbc.co.uk]..

        Reason allows users to specify in advance the decision they want it to reach, and only then to input all the facts. The program's task was to construct a plausible series of logical-sounding steps to connect the premises with the conclusion. The only copy was sold to the US Government for an undisclosed fee.

      • I think this is exactly why Andrew Ng started his machine learning Coursera course, because so many programmers in Silicon Valley were applying machine learning techniques without knowing what they're doing. His idea seems to be, "If I can teach the fundamentals of machine learning to thousands of programmers, then these so-called machine learning 'experts' will be seen for who they are." I hope that managers that think they can be armchair data scientists will also be seen for who they are.
      • by jbolden ( 176878 )

        To build any good model you want a dev set, a test set and a blind sets. Azure should split the data out by default.

  • "so you no longer need to hire someone with a doctorate to use a machine learning algorithm"

    Unless you actually want that someone to actually know what they are doing, e.g., to know that there's no one-size-fits-all "machine learning"...
    • by Anonymous Coward

      If you think of machine learning as being a bit like calvinball, it works out.

    • by Viol8 ( 599362 )

      Not really. Web "developers" don't have to understand TCP/IP to do their thing. I imagine this will be similar. There'll be a basic set of functionality that can be used by Mr Below Average coder to produce ok results.

      • by Anonymous Coward

        Heh, shows what you know :) Mr. Below Average coder will get the Mr. Below Average Result. At least until the computing cost drops so far down that Mr. Below Average MBA will have no problem paying for the huge extra resources needed to blindly stumble through all the preset algorithms (with their corresponding parameter space optimizations) to achieve a still-suboptimal-but-not by-a-really-huge-margin Mr. Average Result. However, there are hard limits for these costs so it's possible the threshold needed f

      • by Trepidity ( 597 )

        More likely, there will be a basic set of functionality that can be used by Mr Below Average coder to generate a bunch of spurious correlations.

        I don't think getting the machine learning to "work" is going to be the hard part, in the literal sense of the code running and generating stuff. But if you have no understanding of statistics, the conclusions you draw are likely to be invalid.

  • If machine learning is such a great thing, why is Microsoft giving it away? It probably hired dozens of machine learning PhDs from top schools, used their insights in its product designs and made a colossal failures out of them. Now their thinking goes, "If machine learning lured us in and fooled us, may be our competitors also would be lured in, made fools of and waste their resources here, and they will pay us a fee to use Azure! win-win!"
    • by pmontra ( 738736 )

      Great things don't have to be secret weapons nobody else can have. Word processors and spreadsheets used to be great things before being given from granted. They fuelled the computer revolution in the 80s (with video games.) Many companies sold them, MS being the most successful in the long run. Same with machine learning frameworks. We'll see how it plays out.

    • If machine learning is such a great thing, why is Microsoft giving it away?

      Because Google, esp. with their investments in DeepMind, are investing in Machine Learning. Giving away what Google's been investing in is a great way to try to encourage lots of start ups to compete with Google's 20% projects.

      These startups will be relying on Microsofts services, not Googles. Which has at least some benefit.

      Besides, they have had a few huge machine learning projects (Dozens of PhDs... they've invested far far mor

  • Microsoft should temper the learning machine cleverness. As someone may ask the machine one day: "which company having a quasi monopoly, is involved in bribery scandals, astroturfing, paid false product reviews and comparisons, ... ?"
  • Could Microsoft be using this newfound machine learning for customer support?

    .
    A month or so ago, I had some issues with Microsoft's Bing bot not following the directions contained in my robots.txt file. When I sent an email in to the BingBot support address, the first reply I got back was that Microsoft considered my robots.txt instructions an "ideal" not something that has to be followed.

    I pushed back and finally got someone who understand the purpose of robots.txt. That person told me to put a work-a

  • This AI machine first has to be able to throw a chair before it can take over Microsoft.

  • HAL: I've just picked up a fault in the AE35 unit. It's going to go 100% failure in 72 hours.

  • I spent sometime at my storage area recently, going through old paper work and history. Not wanting to take anything with me I took photos of the papers of interest as normally do.

    These were legal papers and stuff involving my sons mom when he was much younger. You know the normal break up, the family court, and false accusation, normal stuff.

    Making sure that the photos were saved as they weren't the day before (and 4.2.2) -that night I was updated to (5.0 Lollipop) very nice OS. I had a hard time seeing th

  • PhD in machine learning or ...:

    secretaries - because we can all do our own docs
    car repair mechanics - because it's really just about replacing modules or the whole car
    architects - because there's lots of free 3-D drawing apps out there
    carpenters - because, hey, how hard is it to nail wood together
    lawyers - because just a little reading and memorization will tell you what you need to know
    engineers - because they're like carpenters, only with metal and bigger things
    programmers - because anyone can learn 'hell

You will lose an important tape file.

Working...