By DeVry University
- Dr. Bob Arnot, former Chief Medical Correspondent for NBC
- Dr. Shantanu Bose, DeVry University Provost
In the medical field, artificial intelligence and machine learning are bringing a new set of tools to doctors and healthcare professionals. These resources aid in diagnosing patients, reading images, allocating hospital beds and designing applications with electronic health records.
In this Future-Ready Skills session, Dr. Shantanu Bose and Dr. Bob Arnot give us a glimpse into how these tools are being used to more effectively diagnose malignant melanoma and diabetic retinopathy while helping to increase the accuracy and efficiency of healthcare professionals.
Dr. Shantanu Bose: Hello and welcome to DeVry University's virtual conference, Future-Ready Skills: An Inside Look. I'm Dr. Shantanu Bose, Provost at DeVry University. I'm located in the Chicagoland area, and am honored and delighted to welcome you to today's event.
The upcoming sessions have been created with you in mind. The topics throughout the day are meant to help you think about your own professional and personal future. You will meet with business leaders who are subject matter experts from various industries sharing with you tips to help you handle the current changing norms, questions you should ask yourself now as you consider your career next steps, what working on an agile team looks like, setting goals and overcoming your fears and much more. I hope you can stay the entire day, but if you cannot, no worries. You can jump in and out of the sessions as your day allows. We'll be here until 4:00 P.M. Central Time.
Meet the Presenters
Dr. Shantanu Bose: Among our many experts, you will meet with Groupon’s Chief Technology Officer, John Higginson, who will talk about experimentation and what agile is. You will also hear from two recruiters with Cox Enterprises who will share tips for virtual interviewing. This afternoon, we have a keynote chat with Bob Biglin, CEO at The Center for Advanced Emotional Intelligence. What is emotional intelligence, you may ask? Bob will explain that. You will also meet Alexandra Levit, an author and futurist who will introduce you to the idea of durable value, which is making yourself indispensable in the workforce. The day will close with an inspiring chat with US Olympian, Elana Meyers Taylor, who balances life training and learning. Hear her amazing story and learn how she achieved it all. You may also notice drawings happening in real-time. This is our graphic recorder, Claude, diligently taking graphical notes which will be made available on our website after the event.
Introducing Dr. Bob Arnot
Dr. Shantanu Bose: So, to kick the day off, I have the pleasure of introducing a very special guest, Dr. Bob Arnot. After medical school and training in internal medicine, Dr. Arnot founded the Lake Placid Sports Medicine Laboratory, where he trained Olympic athletes. He was also the medical director for 100+ hospital national emergency services. He was selected first to be on-air co-commentator for ABC's Wide World of Sports and then CBS This Morning with Diane Sawyer. Later, he went on to the CBS Evening News with Dan Rather, then NBC Dateline and the Today Show. He continued on as a war correspondent, joining the first Marine division for the invasion of Iraq and in Afghanistan, and then through the war of terror. He hosted a two seasoned TV show called Doctor Danger and now continues to cover urgent stories like COVID-19 for Fox, PBS, Al Jazeera and Larry King, while undertaking many new projects in deep learning and broadcasting.
One final note before we get started, we'll be taking your questions towards the end of the session so please enter them in the chat window. And now without any further ado, welcome Dr. Bob Arnot.
Dr. Bob Arnot: Shantanu, thank you so much for the gracious introduction. Very, very pleased to be with you.
Getting into Artificial Intelligence and Machine Learning
Dr. Shantanu Bose: Great to have you here too. Thank you again for joining us today. So, to start out, I'm curious, what got you interested in this exciting field of Machine Learning? How did you get started?
Dr. Bob Arnot: Well Shantanu, like a lot of people I kind of looked at artificial intelligence and said not for me – too complicated, too mysterious. I started to poke around a little bit, saw how effective it could be in healthcare and in medicine. And I just started to take some online courses and I said, "This isn't that hard when you dig into it. It's not that difficult. It's incredibly interesting." I have been a data junkie my whole life. There's nothing better than taking a massive data, for instance, with my sports science laboratory.
Dr. Shantanu Bose: I believe that.
Dr. Bob Arnot: For instance, we looked at how people would ski jump, and it turned out that if you pushed off very slowly, you then had a short jump. If you had great angular velocity at your knee, your knee came on court very quickly, you have a very long jump. So, we were able to mine this data out of very sophisticated equipment. We wrote this at MIT, we wrote with various sports teams, and up at my laboratory at Mass General. With this, these tremendous insights jumped out at us, we had a massive dataset, "Hey, jump faster, push faster." And with that, we went from last in the world to close to first.
A lot of people are mystified by data, obviously as a bunch of numbers and rows and columns. It's fantastic in terms of the insights that it gives you, whether it's as an athlete or simple examples. Netflix uses machine learning to figure out what movie you want to watch, which is pretty cool. When you put your various new pictures up on Facebook, it determines who those people are using Machine Learning. So we see it everywhere, but it's just much more accessible than anyone would have ever believed.
The Differences in Artificial Intelligence, Machine Learning, Deep Learning and Data Science
Dr. Shantanu Bose: This is exciting and like you, I'm a bit of a data junkie myself. Love data. So, you use data science. You use the words, Machine Learning. Now you also hear about artificial intelligence that again became popular a while back, and then it has seen a resurgence, right? So, help us demystify these words or these terms, ‘artificial intelligence’, ‘machine learning’, ‘deep learning’ - you have a good background in that - ‘data science’. What is the difference between all of these? Are they all merging and trending towards the same thing?
Dr. Bob Arnot:
- Artificial Intelligence: The term Artificial Intelligence was developed in Dartmouth, which is behind me here, in the 1950s. And I mean, we didn't have the capabilities back then. It was just kind of a fun term to use for a lecture and paper. And I think because we don't have anything that’s truly artificial intelligence. There’s nothing that mimics a human being. All of what we call artificial intelligence it's like the canyon intelligence, one very tiny, narrow little piece. So, when you strip and rip the cover off of artificial intelligence, you really have a couple of things underneath.
- Machine Learning: One is what we would call machine learning. The way you think of this is it takes a whole bunch of data and it figures a pattern out, and that pattern would be what we would call a ‘classifier’. That is, do you have heart disease or don't you, do you have diabetes or don't you, do you have cancer or don't you? So, it's sort of yes or no, which of those do you have? We can go through a couple of examples later. It’s pretty simple to get into, and they're very good tools.
- Data Science: Now you mentioned data science. Anybody who's currently in data science, or interested in it, is set up for success in both machine learning and deep learning, and that is because they understand data structure. Data in medicine health is crazy when you have it from insurance claims, you have it from electronic medical records, you have it from a laboratory, you have it from radiology – all kinds of different data. The data scientist is in an ideal position because they’re able to take and organize that. And if there’s one key thing to walk away from is this, if you can, just organize your data and clean it up, so you don't have missing values or duplicate records, so it's not unbalanced in any way, then you can then use any of these tools. Some of them are off shelf, some of them you don't have to do any programming at all. Now, let me explain that, Shantanu.
- Deep Learning: Let me then go to deep learning. Deep learning is still a lot more mysterious and fascinating. Deep learning basically is what we call these artificial neural networks. The idea is that you would have in the human brain a whole bunch of different neurons, and those neurons would have hundreds of thousands or 10,000 connections between every one of billions of neurons. So, the idea was to try to replicate that. In deep learning, you'll have your input layer here, and then you have a variety of what we call deep layers, which is why they call it deep learning, and all these interconnections between them. And that’s where you learn so much more about data because it's constant looking at the data, trying to figure things out.
So, let me give you an example. One of the neural networks is called a Convolutional Neural Network. And what that does is it analyzes images. And it's been the biggest breakthrough of all in terms of deep learning and artificial intelligence. When you look at these various scans that we would have, for instance, I think there’s something like two billion chest x-rays every year. So how could you learn from that? Well, what you do is you want to take a 100,000 of those x-rays for say lung cancer, and you take the readings that very, very good university professors have of those, so you truly know which ones are positive and which ones are negative. Then you have the Convolutional Neural Network go at it, and it’ll look at every little curve and every little snippet and be able to determine itself what's cancerous and what isn't cancerous.
And with that, then you have the ability to, in a small little hospital like we have up here in Vermont, take a CT scan, an MRI or a chest x-ray, and get as good a diagnosis as you would get at a top university medical center. So wonderful tools, incredibly helpful, very practical, and we see them all day long in everything we use.
Dr. Shantanu Bose: Right. That's great. Thank you for helping us understand those differences. Machine learning predominantly classified data into some segments and answering questions; yes and no type of questions ultimately. Deep learning the system really learns on its own once you give it a ton of data.
How to Get Started in the AI and Data Science Fields with Online Classes
Dr. Shantanu Bose: So, one of the more common questions I get is, “How do I get started in this field?” We have student learners of all ages and some have backgrounds in data science, some do not. Some have a background in coding, some do not. And in anything new sometimes there's fear about “Will I be able to really understand and learn these things?” So how should one think about getting started in this field?
Dr. Bob Arnot: So, here's the advice I give my own kids. I have a 31-year-old, a 25-year-old, and a 7-year-old. With a 25-year-old, especially, all of his friends have gone through college and are kind of looking for the next thing. Or they have friends who are still in college and they have all these ideas that just cause anxiety. They run around with them all day long, like maybe I could do artificial intelligence, maybe I could do deep learning, and they don't ever take any action. They think they have to go and actually sign up for a big university program.
What I advise them to do is to take their smartphone and find a good online course. And we have some terrific ones coming up on healthcare. What's changed in learning is that rather than having to sit down for an hour and a half lecture, many to get immersed in this, in three minutes, someone's going to give you a good little trip introduction, just like you have, right? You're a provost, you understand this. You’ve given people three or four minutes of a very good little introduction.
So learning's changed in that– maybe as you're walking to a bus or subway, or maybe as you're going up an elevator, or instead of going and checking your Instagram, your Facebook, just take three minutes or four minutes or five minutes. It's all what I would call “chunk eyes”. It's a little chunk so you can pick up bits and pieces and mull it over and do the next little piece. And many of these courses will have three or four initial chapters, or I know that you're going to have some totally free courses that will be introductory that people are able to jump in and watch and get a sense of this and just see if they like it. Dig a little bit further.
What I did was I looked at a big introductory course by one of the founders of artificial intelligence, Andrew Ng from Stanford and Google. And I found it very, very theoretical and very complex and I was frankly a little bit put off. Then I found some very good online courses that were just a lot simpler and had these little chunks. From there, your own curiosity takes over. Once I took a bootcamp on artificial intelligence, it's like, well, “You know, linear algebra is really important”. So, I took linear algebra for deep learning. And again, the key with online learning is that anybody can learn anything. It's not like the old days where you had to have four years of some kind of a prerequisite – you can dive in and you can learn.
The reason I really wanted to partner with you, Shantanu, is that you have this dream, this vision, of being able to close the tech gap for the hundreds of thousands of tech gaps that are out there and just make it much more accessible and much more affordable. So, whether somebody that's been laid off during the COVID epidemic or some poor kid in Aleppo, Syria or out in upper Egypt, it gives them this opportunity to be able to earn a certificate, get a job and get started regardless of what their background is. If they're energetic and they're ambitious, they're going to succeed.
Dr. Shantanu Bose: Yeah. You couldn't have said it better. One of the goals and visions that we have is to make education accessible. The tech skills gap is growing. We hear from organizations, employers all the time that they can't hire enough colleagues and employees who are trained. So, making it accessible is absolutely part of our vision.
So just to summarize what you said is:
- Get started. There are plenty of online courses available.
- Start taking these courses in bite-sized chunks. Three minutes, five minutes, ten minutes.
- And then there are so many places to learn these from, right? And if you don't like it, go to the next one. The content is available in many cases.
So that’s good advice.
Dr. Bob Arnot: If you don't like the professor, change. And just to give you a sense of this, many of these courses are like nine dollars. I took up a whole AI bootcamp that was many hours, and they're a very good overview. And I think I spent nine dollars on it. So, it isn't expensive, but it's just your time. Just think about your ordinary day. You don't have to sit down and plug away for an hour at a computer. You can literally walk around with your smartphone, plug in a chapter and get started.
I love the way you said that. And there is just a huge tech gap out there. If you look at jobs, the number one job in America now is data science in terms of salary. You have six-figure salaries for people early on in their careers. There's a huge need all across the spectrum, and it's wonderful, fun, and fascinating as a career. A lot of people, including my own kids, are looking at a pivot during this COVID era. They might have been in the service industry. I have an older son who has been into video production and whatnot, and he's going to do a big pivot just because now is an opportunity. If you're down on your luck, and a lot of us are, you have an opportunity to totally dig in, find something, pivot and really plan for a very bright future because this economy is going to come back and we are going to do well. But get ready for it. Don't sit around worried, anxious and depressed.
There's a phrase, "Don't let a crisis go to waste." And that's very much true. With this kind of a crisis, look around, look for opportunity. It may even be if you are unemployed, that it's a great opportunity for you be able to dig in, find an educational program, sign up, get up, get ready, as the economy comes back and be ready for a new life and a new era in economy 2.0. A lot of people think that economy 1.0 as a vast service economy, wasn’t that great and that 2.0, with the ability to work anywhere, use these Zooms, have all the amazing tool sets on your computer, frees us to live where we want and get much more involved in the information and artificial intelligence era and something we love and that has real legs as a career.
The Impact of Data and Data Production
Dr. Shantanu Bose: Right. There's some statistic you were sharing with me, Dr. Bob, the last time we spoke. It was in a single day we are producing more data than ever before in history? I’d love to hear that statistic again.
Dr. Bob Arnot: It's fantastic. There are a couple of them. One is that we as individuals are putting out about 2.4 gigabytes a day of our own personal data. Now take United Healthcare, with maybe 125 million insurers and do the math on it. It's a lot of data. They're now measuring data in exabytes. If you were to take three or four exabytes, that's a thousand gigabytes. It's so much data that if you have every single word ever spoken in every language, since the beginning of mankind, that would be how much data you have. That's how much we're producing every year in healthcare. It's just way too much data.
The other problem about it is, and you've been very good in terms of data science and how important data science is in this, because it's really the organization of data. And healthcare data is, to be frank, a mess. You look at an electronic medical record. It's not all in nice little columns and rows. It's dictation notes or handwritten notes, or it's an MRI or a lab, or stuff. It's very hard to pull it together. So, they're using natural language processing now to be able to pull out and try to put that into nice rows and columns.
One of the things, as we go along, is to talk about careers here, and if you currently are in data science, or like data, you're going to have a big future even if you don't know all the intricacies of machine learning and deep learning. You're going to be able to prepare that data, look at it, comb through it, size it up, graph different elements of it, see what you think the relationships are, and then pass it onto a machine learning team, or use one of these sites online which are free and allow you to plug and play without coding.
What Makes Good Data?
Dr. Shantanu Bose: Right. And just a follow-up question on data itself. What makes data good data?
Dr. Bob Arnot: It's a great question because it starts, of course, with the accumulation of data. That is, for instance, you can see these smartwatches here, and some of them create great data and some don’t. As an example, heart rate. For anybody who's ever used heart rate when they exercise – just on the wrist, it’s not that great. If you have a chest belt, it's a little bit better data. So, first of all, you have the flawed accumulation of data, so it may be bad quality data.
Another example, here in Vermont, if you had a COVID test, someone would write that down on a piece of paper, they would go to the CDC, they’d put it on another piece of paper, and then somebody would then put it into a computer. Think of all the errors you could make along the way. Look very carefully at how you're accumulating data, how good the sources, how good the tools and techniques are. For instance, Apple Watch now has a way of looking at the most common of all heart arrhythmias, which is called Atrial Fibrillation, and it’s almost as good as a very, very good professional device, but of course you can watch it over a longer period of time.
There are only three key things to look at to make sure it's good, clean data.
- Do you have missing values? And there are techniques in machine learning and deep learning where it can actually take a mean and put a value in there for you if you think it makes sense.
- Second would be, you may have duplicate records which you want to get rid of. Which again is fairly easy to do.
- Then the third thing is you want to balance your data.
One of the things that is so fun about machine learning as you start out, is there are these wonderful libraries. Think of a library as like taking an app off an iPhone or smartphone – and one of them is called Scikit-Learn. With these libraries, you can just graph out the relationships between the data.
So as an example, let's say, you're trying to determine whether or not somebody has congestive heart failure. And in your dataset, you have a thousand people without congestive heart failure, and you have one person with congestive heart failure. Well, it's terribly unbalanced data. There's no way you can get anything out of it. So, you're sizing up the data. Does this make sense? Do we think we have relationships in this data?
And what you'll find is you take just two factors, and that might be, say, blood pressure and blood sugar for diabetes. With heart disease, there’s a pretty tight connection a tight connection between sex and age. You know, a pretty tight connection there. But you're looking for more. So, you may have a hundred columns there and you'll see little relationships there, but that's where the magic of machine learning and deep learning comes in. When people ask about machine learning, I take this example. Shantanu, your parents probably did this. You go out to breakfast and it's like, "Oh my God, I have to spend breakfast with my parents again. What am I going to do?" Well, they bring out this little sheet of paper that has a hundred dots on it. You know, and you go “I wonder what that is.” You start to fill them in and as you're filling it in, at the 70th dot out of a hundred, you go, “That's a bear.”
So, what machine learning does is it’s able to recognize patterns. If there's one concept to walk away with it's this. When I started computer programming, I'm sure when you started coding too, you know, these machine level program, incredibly complex to learn to write thousands of lines of code to get anything out. And this was completely beyond the reach of anybody. The real joy of machine learning is data is writing the computer program for you. So, you take really good, clean, wonderful data, you put it in there, and the machine is writing what we call an algorithm. It’s figuring all the stuff out, it’s writing out an algorithm. So once it's trained up, then you could take your data and put it in.
As an example, let's say you have a prediction of whether somebody has diabetes or not. So you have a training set and that training set basically teaches the computer. The data goes in there and the computer looks at it and molds it. Oh yeah, I see those relationships. And then you would test it. Does this really hold up? Look at the test data, look at the accuracy – being 70, 89, 95, 98%. And then finally, you're ready to use your own data. It's ready to go into action. It's a joy to think you get to just you dump it in and you're sure it's good data. You know what you're looking for there, and it writes the program for you, which is what's transformative over here because coding is hard.
Taking the First Steps: How to Get Started in a Career Related to AI and Healthcare
Dr. Shantanu Bose: Yeah, that's right. We're getting some questions from our audience already, Dr. Arnot. So I'm going to actually go to one of the questions, which was also on my list. So, Carrie B., has asked, “If I want to get started in a career related to AI and healthcare, what would be my first step?”
Dr. Bob Arnot: So, I look at both as, healthcare is a big industry. I don’t know that Carrie can write as we're talking. So, it'd be fun if it's interactive, but I'd ask you, do you have a sense of what you want to do in healthcare? Do you want to do administration? Which DeVry is very strong in. Do you want to be a nurse, do you want to be doctor, or do you want to be a researcher? Because we call this bilingual. I'm a physician, but because now I've learned deep learning, machine learning, I'm bilingual. I have the AI skills and I have the doctor skills. So that's what you're looking at. You're looking, Carrie, at being bilingual.
You want to figure out what is it in healthcare you want to do? You can type that back online if you have an opportunity. Then with the deep learning, machine learning, I would literally go online. There are wonderful sites there. Just search it and it'll pop up. Look at the courses. And once this course is over, go and just pop up the introduction – all of them are three- or four-minute introductions – and go, “Is that interesting?” Look at a couple of free chapters. And again, for $9 or so you can buy an introductory course and dig in and get excited about it. And then once you have, then figure out what are the job requirements?
A new course I'm doing with, Shantanu, is on machine learning. At the end of the course, we have ten real companies that really have jobs listed, that have the jobs listed, and you can look at what the requirements are. Do they want a university degree? Could you have a certificate in it?
They're both wonderful choices because, especially in this COVID era, healthcare turns out to be the most robust of all careers in terms of longevity. If you go into healthcare you probably really do have a job for life, which isn't true for many other industries. And then, as we were saying, with deep learning, machine learning, and artificial intelligence, there's a huge tech gap, a big need, and very, very good salaries to come with it. So, in essence, you've chosen the best two possible careers and put them together in terms of longevity, having a job, interest, and a good steady income.
Dr. Shantanu Bose: Yeah. Then just to summarize health administration, as you mentioned and within that again, health information or health technology. And this is a statistic I remember reading a few years back is a good portion of the healthcare jobs and growing jobs are in the health administration, health information side. So, they’re non-clinical, non-patient facing. Roughly about 40% or so are on the health administration. So definitely, take a look at that. And applications in data science or applications in machine learning, I would presume, would sit on the health information side of healthcare. So definitely a great place to start.
The Future of Healthcare: How Artifical Intelligence (AI) and Machine Learning (ML) Will Change Healthcare
Couple of other questions, while I'm on this chat screen here. Tony S. has asked, “How will AI [artificial intelligence] and ML [machine learning] change the future of healthcare in the next five years?”
Dr. Bob Arnot: It's a good question, because how is it going to change the next three months? I mean, it's moving so unbelievably fast. But to kind of go to the imagination machine, you know, take a Disney-esque look at this. Any hospital worldwide, you're going to be able to get an X-ray and you’re going to get a result that's going to be as good as the best doctor at Harvard or Stanford or Yale. If you are having a stroke and you're in a small little hospital in rural Illinois, you're going to have an instant reading of that as to whether you have a stroke or not, and whether or not you need to be treated. I think the biggest thing for me though, is that we're going to go from a system of catastrophic healthcare, where people have to have heart attacks and have to have strokes and have to develop cancer, to be able to follow them on a millisecond by millisecond basis. Warn them, coax them, coach them, intervene, so there will be less diseases. For example, there's a major university, there’s a picture of mine here, they spent roughly $10 million on an artificial intelligence system and looking at all these variables. And the found that there was a sharp decline in terms of hospitalizations and the utilization of healthcare because they were able to find so much so early.
So, what I would say is that some of the diseases we see now will be a little much less apparent. I think your individual risk is going to decrease. I've actually just written a new book called “Flip the Youth Switch.” And in that book, we looked at a metrical heart rate variability, which you can measure on these watches here. And that actually shows how old you are physiologically. Why is that important? In London, when they looked at patients who ended up on ventilators with COVID, many of them had what they call a black biological age, which meant that they were many, many years past their actual age. Maybe they were 70, but they had the biological age of a 90-year-old.
So, we're going to be able to take and truly reverse aging so that people are physiologically much younger. So, I come out physiologically a 25-year-old and it's a wonderful being able to live life ike a 25-year-old throughout your whole life. It's going to dramatically change the quality of life. We'll have much better interfaces, that is we're going to have much better ways of interacting. So that rather than looking at complex data screens, a huge part of this is going to be a user interface.
I actually took a whole career course in user design just because I think it's the interface that's incredibly important in terms of pulling all this data together – that you can look at a metric on a screen, and it's going to give you a solid piece of advice. I'm a nut when it comes to a user interface, I'll look and if it's not satisfactory, I think people just don't engage. There’s just an interesting aside. So, Nicholas Negroponte at MIT was a great mentor to Steve Jobs and Apple. Interestingly, his background was that of an architect. So, all the computer geeks said, "No, no, no, no, no. It's computation that's going to win. That's what's going to win the battle." And he goes, "No, it's the interface. The interface is going to win." And it did. I mean, just look at our iPhones. We're doing the most complex calculations here, just because we have this interface is up on top of it.
So, interface is going to change, and I would say, too, that's also a great way to enter this space is to understand user design, user experience. In the end, you're looking at how do you convince people to do something? There's no better example that this current COVID era, where you have mass confusion coming out of government and public health and media. People just don't know what to do, wear mask one day, don't wear the next day, wear this kind of mask, wear that kind of mask. It's endlessly confusing because someone isn't sitting back looking at the data. So, to use that exact example, The Lancet, a very famous British medical journal and the oldest of all, looked at 122 different studies, and they found, for sure, if you're using the kind of masks that healthcare providers use, you're getting a 96% protection against the virus. Using the surgical masks, 67%; use a bandana or something, probably closer to 40%; keep yourself three or more feet away, 84% protection; every extra three feet doubles that protection. And if you wore eyeglasses or some sort of glassware, that’s a 78% protection. That's the biophysics. That's it. Very clear, very simple. You just figure out what you're going to do from there.
So when looking at the future of healthcare and machine learning, it's going to be better quality data and a whole lot more of it and from usual places. For instance, we're now looking at depression and anxiety with your smartphone. That is, are you making fewer calls? Are they shorter calls? Are you speaking more slowly when you make the call? Are you moving around less, visiting less, going less places? It's going to be pretty remarkable. It’s going to be really fun, and I think people won't realize they're using it. Just like you don't realize you're using it when you use Netflix or when you throw a picture up on Facebook. It's going to be seamless and in the background.
The Impact of Technology on Healthcare Communication
Dr. Shantanu Bose: Yeah, it is fascinating. Dr. Arnot, we’re getting quite a few questions here, so let's try and see if we can hit as many as we can. This was asked earlier, which goes back to your interface thing a little bit. So, if you could briefly explain this. Carissa H. asks, “How will communication with doctors and healthcare workers change due to advances in technology?”
Dr. Bob Arnot: Carissa, that's a great question. I think the number one change is going to be telemedicine. That means you won't have to leave your job or your home or home office, or if you're taking care of kids, you're going to be able to do almost all of your healthcare via telemedicine. And the change in telemedicine is that, look, you go to a doctor's office because they can examine you. Now you can use telemedicine to be examined by the doctor at home. They have this range of devices. They can look in your ear with an otoscope. They can look into your eye. They can look at your skin with a special device. They can follow your blood pressure, your pulse, your blood sugar. So, they're going to have much more data, and they're going to be able to do that from a remote location.
It also means, I don't know where you live, but let's say you live in rural Louisiana as an example. You're going to be able to have the best doctors at Stanford, Yale or Harvard examine you. A much more democratic system in terms of the ability to use telehealth. So, I think that's big. And the other big thing is again, you're going to be watched on a minute by minute basis so that as insights come up they’ll be able to warn you. You may have an asthma attack, your blood sugar's too low and maybe you should eat some more, or it’s too high and you might need some more insulin. We'll have the automation, for instance, with diabetes that you'll have pumps that are very accurate. They’ll be able to, with an artificial pancreas, treat your disease.
I just think that we're going to have much better overall outcomes. In medicine, we have one word that counts the most, and that is outcome. For instance, I wrote a book 20 years ago called “The Best Medicine.” In that book, we looked at, will you live or die if you have heart surgery? And if it was out in California in a small hospital, there was an 18% chance you would die. If you were at the Cleveland Clinic, there was a 0.9% [chance] you would die. Once these figures were put out there, the hospitals tried very hard to all get online and get better rates. With outcomes, we're going to go across the whole spectrum of healthcare, and you will have vastly improved outcomes across all the chronic diseases. We'll go from below the line, you know here's the disease line here, hopefully pulling up to wellness, gives you much, much, much higher overall level of wellness.
Exploring Business Analytics in Healthcare
Dr. Shantanu Bose: Yeah. Let's take a career related question here. Mark C. asks, “I don't have a healthcare background, but I do have a background in business analytics. Can I go into the healthcare field in AI or machine learning?”
Dr. Bob Arnot: So that's a great question. I would say yes, absolutely. Data is data. With that kind of a background, you'll be able to easily master both machine learning and deep learning. In the courses that I'm doing with Shantanu right now in machine learning and deep learning, we have a whole variety. There isn’t going to be anything very different about it. I mean, clearly, you want to have some kind of expert knowledge you'll be able to accrete, but you’ll be able to very easily fit into a healthcare administration positions as an example. Shantanu, you were saying, I think it’s 40% if healthcare is in some kind of administration. So absolutely, I would jump into this. And use some of our data sets if you want to. You'll be able to load these up, and an understanding of the disease is going to be helpful, but you don't have to be a doctor.
To use an example, with a mammogram. Is it cancerous or isn't it? You don't have to be an expert on all the microbiology, you just need to have a system that is able to accurately predict with very few false negatives or false positives, whether or not this mammogram shows a cancer or not. So, I would say a big yes.
Getting Started with Online Courses
Dr. Shantanu Bose: Big yes, Mark. All right. So, this is a question from Eric Y. And if I'm understanding this correctly, what Eric is asking is, “Where would you spend the next two weeks learning in depth?” So, I'm not sure if that's directed for you specifically who has taken 40 plus courses on this, or is it just in general, “If you had only had two weeks, where would you start?” Which you've already answered, and Eric, if I'm not reading this correctly, please do chime in with a follow up question.
Dr. Bob Arnot: Eric, I would dig in online and just do a search, depending on if it’s automotives and machine learning, or it’s healthcare and machine learning. There are a lot of great sites, like Courser, Udemy, DeVry are very good sites out there. And I would start with the free and inexpensive courses. That is, poke in a little bit, see what you like. If in the first three or four chapters, you like the course, spend $9 or $10. Pay for the course, dig in, and just get excited about it. And, get a variety of courses. You know, you could get four or five chapters in and you could get a little bit stuck. And if you want to take another course, dig in a little bit more. But just dig in, you know, spend time and in two weeks I guarantee you, you're going to be very up to speed on all the technology, all the language, be able to use some real use cases, and decide for yourself that this is for you, and now you want to dig in.
Dr. Shantanu Bose: That's good advice. As little as two weeks or four weeks, you can take plenty of online courses and figure out if:
- This really fascinates you, in which case keep going, or
- It's also good to find out if maybe this is not for you and you can dig in the same in another field.
So good question there.
Exploring Machine Learning in Human Resources
Dr Shantanu Bose: Renee H. asks, “Which AI skills are needed for future success in HR operations?” I'm guessing that's HR in healthcare. Maybe that's where you start, but if you have general comments about HR and artificial intelligence, I would love your thoughts on that?
Dr. Bob Arnot: Renee, I think that machine learning is great for Human Resources. Just on a rudimentary level in terms of being be able to segment through candidates, as an example, looking for potential problems. As a classifier, you're going to be able to classify problems. Now, the interesting thing is you get into the deep learning parts of this, there will be much more sophisticated stuff in terms of how much does somebody socialize? You'll also have all kinds of ethical issues. One very simple one is this, let’s say this. Do you have the right to follow their smartphone? Do you want to see that they went to a bar or a restaurant last night? Did they travel down to Houston, TX over the weekend? So, there will be all kinds of meddlesome problems, but be a whole variety of new and very subtle tools. Especially in the deep learning area, when you look at psychological matches, when you're putting teams together, you may more accurately put those teams together based on compatibility and skills and how complementary they are. So, I think it would be great.
I would start out with the machine learning part of it, just because a lot of this is just going to be classifying people. Then as you want to get to know the more sophisticated aspects of it, you'll be all set once you know machine learning because you're going to know data and how to input the data and what it is. And then I would take your questions, also focus on outcomes. What are your outcomes? Obviously, you want to have better employees and once you have them, you want to keep them happy. As Shantanu does, because I can tell you from everybody I've run into at DeVry, you really want to keep people happy and satisfied.
And polling. A lot of times employees aren't going to tell you if they're unhappy or they have a leader that's angry or making them upset. Just like with illness, we're trying to be preventative and find that illness before it declares itself in an emergency, the same thing with HR. You want to find out, "Hey, here's a team, people are unhappy. They're starting to cast around and what's wrong here? Is it a bad boss?” And unfortunately, there are bad bosses in America. “How can I facilitate reporting so that people don't feel that they're compromising themselves or they ended up some kind of a red line list because they complained?” So, you're going to be able to be much more, bottom line, much more proactive in HR than you currently are.
The Impact of Artificial Intelligence on Pre-med Students and Medical Professionals
Dr. Shantanu Bose: Great. So, let's take a couple of more questions. The questions are coming in fast and furious here. Hodo J. asks, “How will learning this prepare pre-med students to succeed as doctors?” So how will learning this, meaning AI (artificial intelligence), machine learning, prepare pre-med students to succeed as doctors?
I really think that as you apply to medical school, you have to look at an edge. What's your edge going to be? Hopefully have very good grades and you've taken the prerequisites, but is there something there? And I think that the deep learning gives you the “something there.” And in your courses, you're able to apply that you've done some research projects, and you can do those on your own, and then you have something exciting to talk about.
A lot of people when they go to interviews, and you’ll this Shantanu, you more than most here, will stiffen up. It's yes or no, or, “I want to go to university because I...” and they will kind of freeze up. A great way to get to that interview process and to write really compelling essays is to have something you're passionate about. And you will find real passions in artificial intelligence, stuff that you love, and it may also direct you to the kind of career you may find. “Gee, I love image processing, the convolutional neural networks, and medicine, and maybe I want to go into radiology or neuroradiology, or I want to be a neurosurgeon or to be a cardiologist.” So, I think it'll direct you, but it's a wonderful thing to start right now. You can do it in bits and pieces, offline learn Python and learn some of these skills because you have a long time there, and then have something you are really animated and excited about which is going to help you get into medical school and help you be a much better doctor once you're in and through.
Dr. Shantanu Bose: Yeah. I remember, you and I were talking about this. It is becoming increasingly important for doctors to know how to use this. Not know how to code, not know how to program things, but at least know how to use it for a better diagnosis. Learning this at the pre-med stage is a great idea. Good question, Hodo.
PII: How to Maintain Data Security
Dr. Shantanu Bose: There was a question here again from Eric on “How will PII, so that's Personally Identifiable Information, need to change to support collection of data?” That's an interesting one.
Dr. Bob Arnot: It is, because what's happening increasingly is that we have this new kind of area of what we would call precision medicine, that's the kind of technical name for it. In the old days you go, “I have high cholesterol, I'm going to take a cholesterol lowering drug. I have high blood pressure; I'm going to take high blood pressure lowering drug.” Very kind of brute force, very simple stuff. This precision medicine is going to take the greatest intricacies and the greatest data set that there is, and that is the human genome. You'll have your whole human genome. And as you know, that's just an instruction set. What's it doing? So, then you're going to look at the messenger RNA that basically sends out to make stuff happen.
Then you're going to look at, what are you making? Proteins. You are making this particular protein. For instance, in COVID, are you making a tremendous amount of inflammatory proteins? In which case you are at higher risk and may be a candidate for some of these medications like Colchicine early on, or Dexamethasone a little bit further on.
So being able to look at the clinical indicators they currently have, but against the very highly personal background. So, to use COVID again as an example. In Italy they looked at blood types. Nobody had any idea. They were just looking at blood types. But out of this popped this conclusion that if you are a type O blood type you're going to do better, which is very interesting to know. So now, as a physician looking at you and your particular risk and your other risk factors here, I am better suited by knowing your genetic background, knowing your blood type. So I’m just trying to think of what another real-world example would be.
Dr. Shantanu Bose: Let me just interject.
Dr. Bob Arnot: Sure.
Dr. Shantanu Bose: I think the question around PII was more around the data privacy. So, as you're gathering more and more data, are there concerns about data privacy?
Dr. Bob Arnot: There is going to be a whole field around data security, and in medicine, of course, we do have HIPAA requirements. I don’t really see those breached very much. People have the concern and I think there's a career in it, in making sure that this is all highly, highly encrypted – that people using your blockchain, various technologies to be able to segment the data so people can't get at it. I do think as healthcare insurance improves and as the laws improve for human resources, that there will be less of a problem should somebody find out. Well look, we all want our data private. Everybody has stuff that either they are just hyper concerned about or that may cause a real problem – if genetic data came out or you got a particular sensitivity.
So yes, I think that the data security in healthcare is a great place to be because hospitals fret about this every second – that there’s going to be some kind of a data leak. It's a very good question. It's a very good whole area here because you're right, as more and more of this information seeps in – I mean, you having it come from every different source – you have potential for data leaks every which way, maybe even my smartphone – a particular sensor area might give up critical data there, or if I'm sharing it on Strava or another site, it might accidentally give up something that an employer might look at.
As an example, I'm using these training watches. Do I have a heart rhythm disturbance? And could somebody pick that up? And would it be used against me in terms of getting insurance, getting healthcare, getting a new job? So, it's a very insightful question. And the answer is that data security is huge across, obviously, finance and all our other personal information, but of paramount importance, protected by HIPAA. And that's why hospitals pay so much attention to it.
Gaining Practical Experience in the Healthcare Industry
Dr. Shantanu Bose: Right. I'm trying to get to as many people here who have posted questions here. So, let me take one from Darren L. Darren is asking, “How do you get practical experience that is sufficient for working in the industry that is relatively new.” So again, if you're relatively new, I'm focusing this again, back to the healthcare, but let's say you're relatively new to the healthcare industry. How do you get practical experience that is sufficient to land you a role in that industry which is new for you?
Dr. Bob Arnot: So, it's very interesting. I was talking to Google about this, and the interview is no longer, “Do you like to go fishing?” No. The new interview is, “Can you code this for me? See if you can code that problem.”
I ran into the same thing. When I started taking these courses, a lot of them have what they call cookbook code, where you could download the code, put it in, plug and play. The course that I'm doing right now for Shantanu for machine learning actually has that. So, when that's online in the next couple of weeks a month or so, you'll be able to go on and do that. But, what I would say is take as many practical examples as you can. There's a great site called GitHub. That's G-I-T-H-U-B. They have contests on there, so they'll give you a sample problem you could beat. But more importantly, you'll be able to pick up hundreds and hundreds of different examples, which are completely coded data sets. Because I ran into this problem. It was like, "Okay, so I know this stuff and I know how to do it. And I know how to code it. But how do gain any experience?"
So obviously, when you do an internship any place, or preceptorship, or some place you can volunteer. I would jump in and do that at a heartbeat. But on the way there, do use sites like GitHub. And every day take another example, pick it apart, look at how they code it, take it, practice it, upload a site. You'll have lots of practice in terms of uploading data, simple to intermediate to much more complex calculations and computations and scenarios using GitHub. A great site for this.
Final Words of Advice from Dr. Bob Arnot
Dr. Shantanu Bose: Great. So, Dr. Arnot, we are nearing the end of our chat here. Fascinating. We could talk all day. If there was some advice that you would like to leave for our students, for our learners, for those who are already in the field, those who are thinking about healthcare, those who may be intimidated by machine learning, and those who have already had a background in business analytics. We get students of all types. What advice would you have for them? Any parting words here, Dr. Arnot? As we close our session.
Dr. Bob Arnot: What I would say is, as you go to sleep at night, you lie down and you close your eyes, use your imagination. Envision where you could be. That is, envision you could have a much better job, a very interesting job, a very good career, stable, a good salary. Envision all of those. Envision working with smart and interesting people. And there's nothing more fun than working with super smart people – like Shantanu, being super smart. And then, once you have that kind of vision, you have that groundwork and that motivation. Then as I say, dig in, go online, take these very inexpensive courses, get a sense of it, see if you like it or not. And I did that, I was looking at learning artificial intelligence. I looked at a bunch of different other things to do, and I didn't like them. I got in and was like, “That’s not for me, I don’t understand it,” or “That’s not very interesting.” So, see what interests you.
But the big thing you have now you didn't have before is access. You don't have to apply to a university or for a master's or PhD program right away and just figure out you don't like it. You can dig in and find out. But you remember that we're in a very troubling time in terms of careers with this COVID outbreak and people are very worried about it. I'm on every day with my own kids who are very concerned, very anxious. I'm doing a series now with Mass General. The next piece of it is going to be on depression, that it's doubled in this COVID era. So, against that, you really want to take a big step and say, "Look, this is going to be over one day. What could I be?” And really imagine. Don't get caught too much in the practical. Where could you be and how could you develop the skills and just be the best to you ever can?
As a final thought, Shantanu, the old days you go and you have a college degree and that would take you through a 45-year-career. It's not like that anymore. You have to reinvent yourself. And so I have, and I highly recommend having a portfolio still, and it might have user experience in machine learning and healthcare skills. And a portfolio of learning that you're enthusiastic about. Have a portfolio where every year you're learning new skills, because they’re so much more accessible and so much more interesting. It'll make you a more interesting person. A wonderful research paper from Harvard shows that we came out of the cave and made our way into this digital era because there was an intellectual reward in the brain for learning. The more we learned, the more interests we have, the happier we're going to be and the more satisfying life we’re going to have.
Dr. Shantanu Bose: Great parting words here. Become a lifelong learner, have the right mindset, and get started. Thank you, Dr. Arnot, for sharing your interesting and helpful insights into this field of machine learning, artificial intelligence in the area of healthcare. So, thank you again.
Now viewers stay tuned because in a moment you'll meet my colleague, Chris Campbell, who is the CIO at DeVry University, and he'll be speaking to John Higginson from Groupon. So, thank you again for this first segment and Dr. Bob Arnot, thank you again. Have a good day.
Dr. Bob Arnot: Thanks for the great interview and thanks for all you're doing for education and inspiring the next generation to lead great lives and have wonderful careers.
Dr. Shantanu Bose: Wonderful. Thank you and stay tuned everybody.