• Why Do I Keep Babbling about Plastic Surgery, #ai, #machinelearning, #robots, #bigdata ?

    I have recently been on Twitter too much, and I tweet a lot about plastic surgery. I am also fascinated by #ai and #machinelearning, and have had some interesting discussions. In order to clarify some of my tweets and thoughts, I wrote this article.

     

    Why I'm Fascinated by Ai, Machine Learning, Big Data,, and Robots

    Why I’m Fascinated by Ai, Machine Learning, Big Data,, and Robots

     

    What’s AI and Machine Learning and Big Data?
    How is it Being Used in Life Today?
    How Is It Being Used in Medicine Today?
    How Will AI and Machine Learning be Used in Plastic Surgery?
    What Will All of This Mean for Humans and the Purpose of Life?

     

    What’s AI and Machine Learning and Big Data?

    I’m not a computer scientist, I’m a mere plastic surgeon. These are some of my paraphrased definitions of these big terms. I’m not trying to be incredibly precise- I’m just talking in generalizations.

    AI stands for artificial intelligence. I don’t think a world where a computer can talk to you intelligently is that far off. We already have “Siri” on iPhones. No one thinks that Siri is an actual person- but for a freebie on your smartphone, Siri can do lots of cool things and will probably be able to do more in the future.

    Machine Learning is a subset of computer science where you don’t have to tell the software what to do- the software “learns” from lots of data or from experience or both.

    Imagine a software program that has all of the information of the stock market– pricing, the highs, the lows, the price action in the middle of the day, the volume of stocks traded, news stories- and then played a video game with this information. If the software program picks a winning stock, it get a point. If the software program picks a losing stock, it loses a point. You can tell the software program to only gain points, and avoid losing points. With time and lots of data and daily updates of information for each day the stock market is open, the “machine learning” will continue, and the software program will get better and better at picking winning stocks.

    Science fiction? The company already exists- https://www.techtrader.ai

    You can watch their video on Youtube here- https://www.youtube.com/watch?v=KhjLWCokbmY

     

    How is it Being Used in Life Today?

    When you combine AI and Machine Learning with lots of data or “Big Data”, you get a software program that continues to “learn” and try to figure out the best solution for a problem.

    Put the software with a piece of hardware- a robot– and you start to see amazing machines that will only grow in sophistication, power, and faster learning than humans are capable of.

    Amazon bought an entire company – Kiva – so that Amazon could have the best robots in the world to help run their vast warehouses – https://www.youtube.com/watch?v=I-n6fHfUHzA

    We will no longer need human drivers- Google has autonomous cars– http://www.bloomberg.com/news/articles/2015-12-16/google-said-to-make-driverless-cars-an-alphabet-company-in-2016

    We no longer need chefs to make hamburgers– http://www.dailymail.co.uk/sciencetech/article-3670321/Rise-burger-flipping-machines-Robots-churn-400-burgers-hour-set-work-grill-new-restaurant.html

    We no longer need waiters to serve food – customers will order from an iPad, wait for chefs to cook their meal, pick up food from a locker, and eat their food. Oh wait- that’s a restaurant that already exists a couple of blocks from my office – https://www.eatsa.com

    Taxis will all be robots – http://bigstory.ap.org/article/615568b7668b452bbc8d2e2f3e5148e6/worlds-first-self-driving-taxis-debut-singapore

    Large cargo ships with human crews will become relics of the past – http://www.wsj.com/articles/ship-operators-explore-autonomous-sailing-1472635800

    The software program is no longer just directing the hardware- the software is constantly getting fed new data, analyzing it, and adapting to different conditions- resulting in continued learning forever.

     

    How Is It Being Used in Medicine Today?

    To be honest, not much. We do have software that makes medical diagnoses, Every doctor has encountered patients who have already been diagnosed from Dr. Google, Bing, and Yahoo.

    However, carbon based lifeforms with illness, aka “patients” , squirm at the thought of being diagnosed by a machine.

    And the carbon based lifeforms with some medical knowledge,  aka “doctors”,  do not love following algorithms and software.

    We already have the capability of patients describing their symptoms into a website and coming up with a list of possible diagnoses. However, we still expect a doctor to come up with the actual diagnosis and treatment.

    We already have initiatives that comb through patient data and figure out diagnoses faster and more accurately than humans.  We have interesting projects in guiding doctors to make the best treatment options.

    http://www.forbes.com/sites/medidata/2016/01/25/is-machine-learning-the-next-big-thing-in-healthcare/#706bed112f9a

    What happens when the website – with software and lots of data points and tons of patients and constant machine learning- become statistically more accurate than the average human doctor?

    Some types of cancer have treatment options that are experimental, or not common. We already have initiatives where large computer chips have analyzed most or all of charts of cancer patients, and recommend the best chemotherapy, radiation therapy, immunotherapy, or surgery options.

    Artificial intelligence could soon revolutionize the way doctors treat cancer

    What happens when the machine learning software generates a treatment plan that has the best chance of success- and does it in less than 10 seconds?

    In surgery, there are no true autonomous surgery robots linked to Machine Learning. Intuitive Surgical has the “DaVinci” Surgical Robot, but honestly, it’s just a massive cool tool that allows you to do surgery in fairly inaccessible places of the human body. It is guided directly by a human surgeon- it can’t do the surgery autonomously, by itself.

    Would you like to have a fully autonomous surgical robot do your surgical procedure?

    What happens when Version 2.0 or Version 5.0 of that surgical robot is statistically better than a human surgeon? The robot would have fewer complications than a “normal” human surgeon, it never needs to be fed, can work 24/7, can video every move for documentation purposes, and is constantly learning and improving.

     

    How Will AI and Machine Learning be Used in Plastic Surgery?

    We already have software programs that can generate “before and after” imaging. I use the Crisalix software system, and you can alter your face or breast shape and volume to “see” what you would look like with plastic surgery. (https://www.crisalix.com/en)

    We don’t have a surgical robot to do the procedure- yet. For simpler procedures, I definitely see a plastic surgery robot- repairing a simple skin laceration, for example.

    What about cleft lip patients? The cleft lip plastic surgery robot can look at the cleft lip, measure the distances and widths, measure the missing or deficient tissue, compare different types of cleft lip repairs, do the procedure, and watch post-surgery to “learn” what works well and what doesn’t.

    The cleft lip plastic surgery robot can do it again and again with no rest. And it’s constantly learning and evaluating each patient. And it can “learn” to come up with new and different ways to repair a cleft lip- ways that we have never even dreamed about.

    Think of any surgery procedure. The simpler procedures with no scar tissue can have plastic surgery robots do the surgical repair. As robots become more precise and as machine learning continues, the surgical robot will eventually be a “better” surgeon than a human surgeon, with no rest breaks, fewer complications, statistically better outcomes, constant re-evaluation and learning, and never-ending possibilities and options to fix the same problem over time.

    I think that a plastic surgery robot is some time off. I also think that patients will not typically want an autonomous surgical robot to operate on them- yet. I think that any surgical robot will have a very hard time with scar tissue and complicated cases – but not for long.

     

    What Will All of This Mean for Humans and the Purpose of Life?

    Ahh, this is where I get annoying and provocative on Twitter.

    In the near future, AI robots will create jobs for the humans that create robots and repair robots. Also, anything related to the basics of life- food, shelter, clothing, healthcare- will also continue to create jobs.

    What happens when the robots become more sophisticated? Almost every job is up for grabs.

    Farmer that plants? There’s a robot for that- https://www.youtube.com/watch?v=8r0CiLBM1o8

    A harvesting machine? There are robots that as well. In fact, there will be a robot farm in 2017 – http://www.dailymail.co.uk/sciencetech/article-3264858/World-s-robot-FARM-open-2017-Firm-plans-replace-humans-machines-lettuce-factories.html

    There are already robots that can lay bricks – and in the future, make shelter – https://www.youtube.com/watch?v=MVWayhNpHr0

    There will be robots – that can make and assemble even more robots.

    There will be robots that will be able to repair- other robots.

    There will be robots that will have machine learning- and create new robots that have never been seen before that will be more efficient.

    Will there be a day when robots can do everything that we need? Create music? Create art? Diagnose our illnesses? Come up with the best treatment options? Do surgery on us? Is Skynet in the movie series “Terminator” a real threat? Will Ai and Machine Learning ever actually have a conscience and achieve self-actualization?

    Maybe Version 1.0 of many of these robots will not be that wonderful and can’t work as well as a human. What about Version 2.0? Version 19.0?

    More advanced versions will be better than humans in almost every field of work.

    I believe that many jobs will be taken over by robots very soon. I think my job is safe for a slightly longer period of time- but yes, someday sooner than you think, a robot will do your job better than you, and a plastic surgery robot will be doing my job.

    Any comments? Questions? Want to share this article? Feel free to fire away below!

    Leave a reply →
  • Posted by Minas Chrysopoulo MD on September 3, 2016, 11:54 am

    Great article Roy. It’s also great to see artificial intelligence being applied to cancer diagnosis and treatment. IBM’s Watson is a great example and an incredible tool. I don’t know about you, but I can’t read, evaluate and remember hundreds of thousands of pages of medical journals in just a few seconds. How about evaluating trends in huge volumes of anonymous medical data, interpreting pathology slides and radiology images more accurately than human physicians? These are examples of technology that is already being tested. At some point we could have routine access to these tools. How much better could we become as physicians?

    Reply →
  • Posted by María Jesús Díaz on September 3, 2016, 9:38 pm

    Great post! And amazing blog!
    I am looking forward to see AI in our hospital lifes, specially helping with the more boring and complicated radiologists tasks.
    Analizing complex internal medicine or oncology medical records will be easier (not so important to surgeons I guess).
    Robotic surgical techniques are many times oversold.
    I think the life robotization will make the surgeons and specially PS both technical and craftsmanship work increasingly valued.
    Congrats 😉

    Reply →

Leave a reply

Cancel reply

Photostream