Slashdot Mirror

← Back to Stories (view on slashdot.org)

Microsoft To Launch Machine Learning Service

Posted by Unknown on from the point-and-click-your-way-to-ai-winter dept.

angry tapir (1463043) writes Microsoft will soon offer a service aimed at making machine-learning technology more widely usable. "We want to bring machine learning to many more people," Eron Kelly, Microsoft corporate vice president and director SQL Server marketing, said of Microsoft Azure Machine Learning, due to be launched in beta form in July. "The line of business owners and the marketing teams really want to use data to get ahead, but data volumes are getting so large that it is difficult for businesses to sift through it all," Kelly said. The service will have "...an interface called the Machine Learning Studio. The palette includes visual icons for some of the most commonly used machine-learning algorithms, allowing the user to drag and drop them into a visually depicted workflow." Algorithms themselves are implemented in R, which the user of the service can use directly as well.

56 comments

  1. AC to Launch First Post by Anonymous Coward · · Score: 0

    get yourself a frosty bitches!

  2. Sky.NET by Anonymous Coward · · Score: 0, Funny

    Why did it have to be Microsoft?

    1. Re:Sky.NET by sribe · · Score: 4, Funny

      Why did it have to be Microsoft?

      Well, at least it will keep crashing, and maybe humanity will figure out a way to defeat it during one of the reboots.

    2. Re:Sky.NET by Cryacin · · Score: 4, Funny

      (Arnie) Reboot this!

      --
      Science advances one funeral at a time- Max Planck
    3. Re:Sky.NET by AHuxley · · Score: 1

      It only crashes for the end user, the parts that track the user for advertizing and the US gov have a perfect operational record.

      --
      Domestic spying is now "Benign Information Gathering"
    4. Re:Sky.NET by Anonymous Coward · · Score: 0

      Why did it have to be Microsoft?

      Duh. Who else has the extended history of corporate malevolence and spectacular incompetence required to create an AI that dooms all of humanity?

    5. Re:Sky.NET by Anonymous Coward · · Score: 1

      Well, at least it will keep crashing, and maybe humanity will figure out a way to defeat it during one of the reboots.

      Sadly, they appear to be embracing R (http://www.r-project.org/) as the backend, so it will be stable ate least until the "extend" phase.

    6. Re:Sky.NET by Anonymous Coward · · Score: 0

      In the words of Shawn Hargreaves on an XNA presentation that touched on reflection: "if you want to build sky net and take over the world, this is how you do it". On a more serious note I can't really picture a machine learning service being all that popular compared to a machine learning API. When it comes to machine learning you need fine grain control for your fitness tests. Writing a good fitness test for all but the simplest of problems usually involves weeks, months and even years on end of writing code. Imagine if you had to give up 100+ WPM typing for drag and drop, productivity be damned.

    7. Re:Sky.NET by Jon+Peterson · · Score: 2

      Google, Apple, Oracle, IBM, etc. etc.

      Actually, Azure is great, and the addition of high level services like this is the right direction. Just spinning up VMs isn't nearly as useful as a service layer.

      The algorithms aren't an especially hard part of machine learning, dealing with the data is. Anything that would save me the hassle of trying to fit things in RAM would be great...

      --
      ----- .sig: file not found
    8. Re:Sky.NET by LifesABeach · · Score: 2

      I guess we will see R# in the next version of Visual Studio?

    9. Re:Sky.NET by Anonymous Coward · · Score: 0

      Fuckwit.

    10. Re:Sky.NET by RevWaldo · · Score: 1

      It looks like you'd like to kill all humans.

      [ ] Launch thermonuclear weapons at all the major population centers.

      [ ] Don't help, and let me target the weapons directly.

      [ ] Don't show this tip again.

      .

      --
      Prisencolinensinainciusol. Ol Rait!
    11. Re:Sky.NET by Anonymous Coward · · Score: 0

      typical unformed /. myopic rubbish

  3. You want machine learning? You want IBM's Watson. by mmell · · Score: 4, Interesting
    That machine wasn't merely a decision engine or a huge database - it was a learning system which was given a few months to learn all it could on a broad variety of subjects by crawling the internet. It wasn't programmed with rules to pair an answer with the correct question; it was a system which "learned" how to associate concepts. It was programmed to permit it to weigh its questions against how well they correlated to the answers - to determine a confidence level - but it wasn't specifically programmed to devise questions associated with answers. That's why the next publicly stated idea for the system was in medical diagnostics. It's another area where the ability to relate multiple seemingly disparate items of information with a non-static data store seemed to be of value.

    Giving the right programming, it might even hold a conversation better than a 13 year old Ukrainian boy.

  4. Give WEKA a try by AlienSexist · · Score: 2

    WEKA is Open Source, has an adequate GUI, many different kinds of algorithms available, and a "knowledge flow" visual designer for you to chain it all together. I've used it in a few personal and professional projects to find things like which variables most strongly influence an outcome, decision trees, derived formulas and expressions that accurately predict outputs from inputs, and various kinds of data visualizations for clustering data samples. Code is in Java so I presume you could embed it within a system to automatically perform analysis and swap algorithms on the fly. Best of all, since this is software under your control, and not a Corporate-offered service, your valuable data never leaves your control.

    I think WEKA already did a lot to make these kinds of data analysis accessible as Microsoft is aiming to do. No matter who provides it to you, there is something totally awesome about being able to click a few buttons and get some interesting results to munch on.

    1. Re:Give WEKA a try by lorinc · · Score: 4, Insightful

      I have only one problem with fancy GUI that allow you to train a predicting model in 2 clicks: how confident can you be in your model, since all the parameters are masked and you have no knowledge about them? I still think it is dangerous to rely on a tool you don't understand and you can't control up to a satisfactory level, especially when it is to be used in prediction - something we expect to be highly reliable in many aspects due to old development of science like balistics.

      I've written a ML library myself (also in Java, more lightweight than weka, but with no gui - although it comes with standalone binaries for some basic setups) and I can tell you there is no good default tunning that works well for every kind of situation. ML is seriously a young science that gets rapidly tricky even on very common problems, which is very different than field for which we have very accurate solvers that work most of the time (again balistics is probably a good example, at least because it is taught in school and sets the prototype of what we name science). I fear hidding the youth (and thus the imperfection) is only going to cause damage through misconception and false interpretation.

      --
      Video of some good progressive thrash music
    2. Re:Give WEKA a try by LetterRip · · Score: 4, Informative

      Having used Weka, sklearn, vowpal wabbit, Orange, and lots of others - I'd say that Weka is my least favorite, it is slow and has crappy memory management. I do use it occassionaly for feature ranking and reduction, but rarely for the actual machine learning.

    3. Re:Give WEKA a try by Anonymous Coward · · Score: 1

      I actually worked on this project and I can say from using it and WEKA that this is much, much more user friendly.

    4. Re:Give WEKA a try by Anonymous Coward · · Score: 0
      Translation:

      I got money from this and I want more.

    5. Re:Give WEKA a try by Anonymous Coward · · Score: 0

      Was good for learning machine learning.

      But being the skeptical CS I am, I don't trust anything I haven't built. Not because of the construction. But how can I understand that which I have not built?

      Eg., the human is the error in this puzzle. Frail human mind, with all its overconfidence and assumptions :-(

    6. Re:Give WEKA a try by Jon+Peterson · · Score: 3, Insightful

      You don't trust anything you haven't built???

      How do you know your HDD firmware isn't corrupting data? Build it yourself??

      --
      ----- .sig: file not found
    7. Re:Give WEKA a try by Anonymous Coward · · Score: 0

      Having used IKEA, skream, vocal vomit, Orange, and lots of others - I'd say that IKEA is my least favorite, it is slow to assemble and uses crappy much room for very small storage space. I do use it ocassionaly for laundry drying and empty space reduction, but rarely learn from its actual machinery.

    8. Re:Give WEKA a try by Anonymous Coward · · Score: 0

      What's your favorite system? Very curious.

    9. Re:Give WEKA a try by snowsmann · · Score: 0

      Please mod this funny! This sentence is hilariously well made.

      --
      timeo Danaos, et dona ferentis
    10. Re:Give WEKA a try by Anonymous Coward · · Score: 0

      Write your own unit tests and test it, duh !

    11. Re:Give WEKA a try by radtea · · Score: 2

      I have only one problem with fancy GUI that allow you to train a predicting model in 2 clicks: how confident can you be in your model, since all the parameters are masked and you have no knowledge about them

      This. Expecting people with no knowledge of statistics, programming, or machine learning algorithms to develop appropriate models and interpret them correctly is unrealistic.

      I've worked on applications that attempted to bring machine learning to the desktop for specific tasks, and it failed because naive users were simply not able to bring the required knowledge and nuance to the table. It made experts better (I still use the program myself now and then) but there was a threshold below which it was useless.

      Without getting your hands on the data and understanding something about the internals of the problem you can't build learning systems that will be plausibly generalizable. Overtraining (and undervalidation) is still a rampant problem in ML applications, and this kind of technology will just make it worse.

      Learning algorithms, especially at the level they are implemented in R, are not black boxes. They need to be understood and applied carefully and appropriately.

      --
      Blasphemy is a human right. Blasphemophobia kills.
  5. Watch out for embrace and extend by Anonymous Coward · · Score: 1, Insightful

    The R programming language is fully GNU. And now microsoft wants it. Expect them to take it, then try to claim it, and then when they can't, try to make their own, and drag everyone away from it (bringing a world of hurt to all involved).

    1. Re:Watch out for embrace and extend by Anonymous Coward · · Score: 0

      microsoft is like cancer to use their words.

    2. Re:Watch out for embrace and extend by shutdown+-p+now · · Score: 1

      Microsoft has been doing Python for over 3 years now. A lot of people "expected them to take it, then try to claim it, and then when they can't, try to make their own, and drag everyone away from it", but it hasn't materialized yet.

      Have you considered that your preconceived notions of Microsoft are a decade old, and it's might be a different company run by different people by now? Or that there's no business purpose in anything beyond "embrace", if you can just use it as a selling point to sell a product to people who don't care about all-MS top-to-bottom stack, as has been increasingly common in the last ten years?

  6. Re: You want machine learning? You want IBM's Wat by Anonymous Coward · · Score: 0

    .. yet IBM has failed miserably in trying to productize it. IBM's latest attempt is an open competition for mobile developers to find a way to productize it for them.

  7. Re: You want machine learning? You want IBM's Wat by Anonymous Coward · · Score: 0

    .. yet IBM has failed miserably in trying to productize it. IBM's latest attempt is an open competition for mobile developers to find a way to productize it for them.

    Microsoft's doing the opposite.

    Nobody wants Cortana or Windows Phones, so they're trying to monetorize the machine learning backend.

    Judging by experience though, this will be to machine learning what SharePoint is to web services. Stick with PredictionIO or Mahout if you value your sanity...

  8. Leave Windows OS as first step forward by ikhider · · Score: 0

    Machines realizes nothing gets done with the Windows OS; too much bloatware, too many backdoors and defrag/antivirus required to run the OS--immediately switches to free, libre software and never looks back.

    --
    "SO we bide our time, waiting for a purer kick to bloom and the future is still bleak, uncertain and beautiful" -GSYBE
    1. Re:Leave Windows OS as first step forward by Anonymous Coward · · Score: 0

      defrag

      Lame troll, bro. Defrag has been unnecessary for the past 15 years.

    2. Re:Leave Windows OS as first step forward by Anonymous Coward · · Score: 0

      You don't want to start a flame war, do you?
      Manual defrag is unnecessary. MS has only automated the click onto the defrag button.

    3. Re:Leave Windows OS as first step forward by FearTheDonut · · Score: 1

      I believe defragging is completely automated at this point (since Windows 7) - no button press needed (but you can do it, if you want too).

  9. When have they ever implemented this well? by Anonymous Coward · · Score: 1

    I mean, they have a lot of products and very few, if any, are examples of good machine learning.

    1. Re:When have they ever implemented this well? by Anonymous Coward · · Score: 1

      Clippy.

    2. Re:When have they ever implemented this well? by hoggoth · · Score: 1

      Microsoft has finally gone too far:

      https://flic.kr/p/nJ4xag

      --
      - For the complete works of Shakespeare: cat /dev/random (may take some time)
  10. Re:You want machine learning? You want IBM's Watso by skovnymfe · · Score: 1

    Wouldn't you have to ask it questions in the form of an answer? But I suppose that's what diagnosticians do now I think about it.

    Six patients, one of which presents with a big hole in the head.

    What is russian roulette?

  11. Re: You want machine learning? You want IBM's Wat by TapeCutter · · Score: 1

    AFAIK, IBM has only just started leasing instances of watson to "development partners" earlier this year. A single instance of Watson can now run on a bar-fridge server, I doubt they come cheap and I doubt IBM will simply allow people to access/deploy the technology without IBM's "help". They have a clear market advantage and a keen interest from the big end of town, they are not going to throw that away to compete with a MS gimmick.

    --
    And did you exchange a walk on part in the war for a lead role in a cage? - Pink Floyd.
  12. Journalists by louic · · Score: 1

    Let's hope journalists will not be told about this tool. I'm getting tired of repeating that correlation does not imply causation.

  13. Microsoft Services by rossdee · · Score: 1

    They advertise that the Microsoft Cloud gives the Lotus F1 Team the winning edge

    Lotus haven't done much winning this season, they are the eighth team out of 11 in the constructors championship

    1. Re:Microsoft Services by Richard_at_work · · Score: 1

      Microsoft also partner with McLaren F1 to produce the engine ECU :)

  14. Triple E by fph+il+quozientatore · · Score: 2

    So, R is being Embraced. I wander what's coming next...

    --
    My first program:

    Hell Segmentation fault

  15. Re:Seriously, fuck those people! by Anonymous Coward · · Score: 0

    Why this is modded down?

  16. No one will know by Viol8 · · Score: 1

    Nobody watches F1 anymore anyway. Its just an out of touch travelling fashion show with cars these days with utterly boring racing and overpaid drivers who've all had a personality bypass. Eventually it'll disappear up its own exhaust pipe with barely a squeak of tyres and nobody will even notice.

  17. Do they know KNIME? by Anonymous Coward · · Score: 0

    "The palette includes visual icons for some of the most commonly used machine-learning algorithms, allowing the user to drag and drop them into a visually depicted workflow"

    This feature was surely inspired by the open-source KNIME: http://www.knime.org/knime

  18. Re:Seriously, fuck those people! by Anonymous Coward · · Score: 0

    When they had, it wouldn't have been so successful. They get the money for not touching it.

  19. You do know what's coming, right? by ArcadeMan · · Score: 1

    We keep getting news about automated weapons, cheaper and cheaper drones, and now machine learning.

    Everyone saw the Terminator movies, right?

    --
    Get free satoshi (Bitcoin) and Dogecoins
  20. The idea is like RapidMiner by xgdgsc · · Score: 1

    Where you can draw a visual workflow.

  21. New recipe? by Anonymous Coward · · Score: 0

    - Pick some random (and awesome) piece of free software

    - Slap some turd on top of it

    - Profit?

    Seen that pattern 'couple of things too often (git + atlassian => stash, my company buys that because... enterprise, ho, ho).

    Now contributing to some cool free project would be something quite different.

  22. Re: You want machine learning? You want IBM's Wat by Anonymous Coward · · Score: 0

    Are you sure, most of Watson is OpenSource and there was an article on how to build one in Forbes as well as one of the Watson engineer's IBM blogs.