The Digital Twin

  • Comments posted to this topic are about the item The Digital Twin

  • I've worked on a big personalisation project that produced a distance score between 250,000 products for 5 million customers.  That was 1.25 trillion records generating 37TB of scoring data.
    The capabilities of the team working on the project and the capabilities of the prototype system were far in advance of the imagination of the key decision makers.  The technical hurdles were the easy part.

    It took a significant amount of work to sell the idea to stakeholders and, God bless them they rose to the challenge.  Ironically the project was killed by fear of the unknown in some sections of the IT department and not by lack of faith from the commercial teams.


    Preventing people seeing adverts for products they have already seen can be done using HyperLogLog to estimate whether a customer has already seen that product.  
    HyperLogLog bears some studying.

    • Good for estimating membership of a set
    • Good for estimating count distinct
    • Gives truly insane rates of data compression
    • May give false positives but won't give false negatives 
    • Accurate to around 2%.  2% error rate is fine for personalisation.


    AWS RedShift already uses HyperLogLog when you prefix a count distinct with the approximate keywordselect approximate count(distinct pricepaid) from sales;

    In short, the technology and capability put forward in your editorial already exists.

  • ... This personalization has an annoying side as well, after all many of us have experienced more targeted advertisements and annoying uses of our personal information, however much of that is crude, and lacking in sophistication. Perhaps if the models using data about me would be less annoying if companies didn't try to sell me a laptop a week after I've purchased one, or show me sales on products when I'm looking at SQL Server articles. ...


    The advertisers would argue the problem is that we're just not sharing enough of our personal information with them. If we would simply open the flood gate and let them in on every intimate detail of our personal life, they could place that add for Cialis in font of us the moment we need it most.

    "Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho

  • I'm sorry Steve, but I have a very hard time grasping this idea of "digital twins". All I could come up with is maybe cloning a database, or developing a template of a Visual Studio project so you could create a new version of that project. For me, the word, "twin" means it is of the same substance or origin (i.e.: the same parents and siblings to one another, born at the same time, with identically shared DNA). And even the link in the article didn't help. That's more likely, I'm guessing, because the video is hosted on some social network. My employer blocks all social networks.

    I Binged it and came up with this link instead.

    I realize that it isn't your fault the adopted phrase is "digital twin". It wasn't your idea. Whoever came up with it, IMO, came up with a bad and confusing phrase.

    Kindest Regards, Rod Connect with me on LinkedIn.

  • Rod at work - Wednesday, July 19, 2017 8:34 AM

    I'm sorry Steve, but I have a very hard time grasping this idea of "digital twins". All I could come up with is maybe cloning a database, or developing a template of a Visual Studio project so you could create a new version of that project. For me, the word, "twin" means it is of the same substance or origin (i.e.: the same parents and siblings to one another, born at the same time, with identically shared DNA). And even the link in the article didn't help. That's more likely, I'm guessing, because the video is hosted on some social network. My employer blocks all social networks.

    I Binged it and came up with this link instead.

    I realize that it isn't your fault the adopted phrase is "digital twin". It wasn't your idea. Whoever came up with it, IMO, came up with a bad and confusing phrase.

    The digital twin is a model of how something, say a jet engine, performs. Using data from the engine, combined with flight patterns, weather, cameras, etc., the model specifically outputs the probabilities of maintenance needed or potential faliures. They used to use a model for all engines, but an engine flying in the US Pacific Northwest experiences much different wear than one flying in the middle east. They found that with a single model, too much maintenance on the former, not enough on the latter.

    There is a platform, where a model  for each  jet engine in service is constantly running and  updated with data. This is the  "twin", personalized to that specific engine.

    They are doing this with many of their products. Just as some doctors are modeling a drug (or multiple drugs) specific behavior in your body, not using the results of a some study.

  • Steve Jones - SSC Editor - Wednesday, July 19, 2017 9:12 AM

    The digital twin is a model of how something, say a jet engine, performs. Using data from the engine, combined with flight patterns, weather, cameras, etc., the model specifically outputs the probabilities of maintenance needed or potential faliures. They used to use a model for all engines, but an engine flying in the US Pacific Northwest experiences much different wear than one flying in the middle east. They found that with a single model, too much maintenance on the former, not enough on the latter.

    There is a platform, where a model  for each  jet engine in service is constantly running and  updated with data. This is the  "twin", personalized to that specific engine.

    They are doing this with many of their products. Just as some doctors are modeling a drug (or multiple drugs) specific behavior in your body, not using the results of a some study.

    This has been around for years in different forms, indeed some situations may be better off with analog twins. Many years ago I remember an article about a flood model for the Mississippi river. It was only a few hundred feet long but contained all the major tributaries. By mimicing the reported rain distribution in the different areas of the system, and pouring water into the tributaries in appropriate sequence (because the model is smaller it ran on a much faster time scale), they could predict flood stage heights and times along the river. To model the Mississippi digitally requires a substantial supercomputer.

    [Now, thinking about the word 'analog', we've become accustomed to think 'non digital', but the original word meant 'something substantially similar' to the item in question, as in analogy]

    ...

    -- FORTRAN manual for Xerox Computers --

  • jay-h - Wednesday, July 19, 2017 9:39 AM

    This has been around for years in different forms, indeed some situations may be better off with analog twins. Many years ago I remember an article about a flood model for the Mississippi river. It was only a few hundred feet long but contained all the major tributaries. By mimicing the reported rain distribution in the different areas of the system, and pouring water into the tributaries in appropriate sequence (because the model is smaller it ran on a much faster time scale), they could predict flood stage heights and times along the river. To model the Mississippi digitally requires a substantial supercomputer.

    [Now, thinking about the word 'analog', we've become accustomed to think 'non digital', but the original word meant 'something substantially similar' to the item in question, as in analogy]

    It has, but there hasn't been the ability to do this efficiently at scale. Part of what GE did was like what MS does with Azure, Amazon with AWS, they made a platform so they can apply many models to each instance. This would be like someone modeling various sections of the Mississippi with physical models.

  • There really is no technical concept that hasn't already been conceptualized in an episode of Star Trek.

    "Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho

  • Eric M Russell - Wednesday, July 19, 2017 12:55 PM

    There really is no technical concept that hasn't already been conceptualized in an episode of Star Trek.

    Star Trek didn't have the idea of a smart phone and social apps like Twitter and Facebook. It's technology, economy, and social structure were Unobtainium, Handwaveium and Utopian. And we still don't have flying cars or shuttle craft or warp drive or transporters.  😛

    Digital twins are good for modeling small, well understood sections of the physical world, like engines, structures and electronics. They fail when applied to complex chaotic systems like climate, economy or society.

  • chrisn-585491 - Thursday, July 20, 2017 7:03 AM

    Eric M Russell - Wednesday, July 19, 2017 12:55 PM

    There really is no technical concept that hasn't already been conceptualized in an episode of Star Trek.

    Star Trek didn't have the idea of a smart phone and social apps like Twitter and Facebook. It's technology, economy, and social structure were Unobtainium, Handwaveium and Utopian. And we still don't have flying cars or shuttle craft or warp drive or transporters.  😛

    Digital twins are good for modeling small, well understood sections of the physical world, like engines, structures and electronics. They fail when applied to complex chaotic systems like climate, economy or society.

    In fact, Star Trek totally did feature the equivalent of smart phones; a device they called the communicator. In STNG, it was more than just a communicator but was a voice input interface to the ship's AI computer.

    There was also at least one STNG episode where the engineering team created a virtual holographic representation (digital twin) of the starship's warp drive so they could test different configuration scenarios.

    ... La Forge realizes that the only way to restore power is to reconfigure the warp drive and traces the warp drive's design back to the Enterprise's construction and blueprints created by Dr. Leah Brahms. La Forge enters the ship's holodeck to help figure through the engine reconfiguration, whereupon the computer takes an off the cuff remark by Geordi literally and creates a holographic representation of Brahms herself to assist him in his work. As Geordi does so, he asks the computer to update the hologram with Dr. Brahms' personality profile, and ...


    https://en.wikipedia.org/wiki/Booby_Trap_(Star_Trek:_The_Next_Generation)

    This cartoon says it all.
    Image result for onion cartoon steve jobs iphone

    "Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho

  • Eric M Russell - Thursday, July 20, 2017 7:29 AM

    chrisn-585491 - Thursday, July 20, 2017 7:03 AM

    Eric M Russell - Wednesday, July 19, 2017 12:55 PM

    There really is no technical concept that hasn't already been conceptualized in an episode of Star Trek.

    Star Trek didn't have the idea of a smart phone and social apps like Twitter and Facebook. It's technology, economy, and social structure were Unobtainium, Handwaveium and Utopian. And we still don't have flying cars or shuttle craft or warp drive or transporters.  😛

    Digital twins are good for modeling small, well understood sections of the physical world, like engines, structures and electronics. They fail when applied to complex chaotic systems like climate, economy or society.

    In fact, Star Trek totally did feature the equivalent of smart phones; a device they called the communicator. In STNG, it was more than just a communicator but was a voice input interface to the ship's AI computer.

    There was also at least one STNG episode where the engineering team created a virtual holographic representation (digital twin) of the starship's warp drive so they could test different configuration scenarios.

    ... La Forge realizes that the only way to restore power is to reconfigure the warp drive and traces the warp drive's design back to the Enterprise's construction and blueprints created by Dr. Leah Brahms. La Forge enters the ship's holodeck to help figure through the engine reconfiguration, whereupon the computer takes an off the cuff remark by Geordi literally and creates a holographic representation of Brahms herself to assist him in his work. As Geordi does so, he asks the computer to update the hologram with Dr. Brahms' personality profile, and ...


    https://en.wikipedia.org/wiki/Booby_Trap_(Star_Trek:_The_Next_Generation)

    This cartoon says it all.
    Image result for onion cartoon steve jobs iphone

    The communicator was nothing more than a scaled-down radio, not a true smart phone. Star Trek overlooked personal computers, the ships computer was just a mainframe, the tricorders were never fully explained unless the plot required some magic. AI was in literature and industry journals long before Star Trek. The aerospace industry has been using simulations and modeling for decades.

    There really isn't any Star Trek concept that was already in progress or literature (fact or fiction)  before Star Trek. Being a television show, it's gotten more exposure and may have inspired a broader audience than others, but it's a drama with a lot of fantasy.

    As much as science fiction has been inspirational, it's entertainment and escape fiction. Once you plunge into the nitty-gritty of science and math, its a awful lot of hard work with a beauty of truth and confusion. The problem is that people confuse the entertainment for the science and don't respect the work.

  • Maybe I'm easily wowed, but from MS Build in the spring, I was very impressed with the idea of digital twins, including this video:

    https://www.youtube.com/watch?v=GAkB98ewcjI

    Trying to figure out the world of SQL as marketing consultant for SQL Solutions Group https://sqlsolutionsgroup.com/

  • David.Poole wrote:

    I've worked on a big personalisation project that produced a distance score between 250,000 products for 5 million customers.  That was 1.25 trillion records generating 37TB of scoring data.

    The capabilities of the team working on the project and the capabilities of the prototype system were far in advance of the imagination of the key decision makers.  The technical hurdles were the easy part.

    It took a significant amount of work to sell the idea to stakeholders and, God bless them they rose to the challenge.  Ironically the project was killed by fear of the unknown in some sections of the IT department and not by lack of faith from the commercial teams.

    Preventing people seeing adverts for products they have already seen can be done using HyperLogLog to estimate whether a customer has already seen that product.

    HyperLogLog bears some studying.

    • Good for estimating membership of a set
    • Good for estimating count distinct
    • Gives truly insane rates of data compression
    • May give false positives but won't give false negatives
    • Accurate to around 2%.  2% error rate is fine for personalisation.

    AWS RedShift already uses HyperLogLog when you prefix a count distinct with the approximate keywordselect approximate count(distinct pricepaid) from sales;

    In short, the technology and capability put forward in your editorial already exists.

    I'm missing it... what would have been the big advantage if the project had continued to a successful conclusion?  Simply not showing someone the same ad more than x number of times?

    --Jeff Moden


    RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
    First step towards the paradigm shift of writing Set Based code:
    ________Stop thinking about what you want to do to a ROW... think, instead, of what you want to do to a COLUMN.

    Change is inevitable... Change for the better is not.


    Helpful Links:
    How to post code problems
    How to Post Performance Problems
    Create a Tally Function (fnTally)

  • This was removed by the editor as SPAM

Viewing 14 posts - 1 through 13 (of 13 total)

You must be logged in to reply to this topic. Login to reply