My Gosh! ‘the talks’. Everyone is talking about ‘AI’. Everyone is talking about ‘AI’ like everyone is using ‘AI’. Trust me, there is no shortage of ‘Talk’, so we are going to hop on the band wagon and talk about ‘AI’ too (seems timely) and what it means to the world of Healthcare.
Artificial Intelligence ‘AI’ is a very (very) broad spectrum under which lies several sub categories of techniques that are used to realize ‘AI’. For the benefit of all readers, let’s breakdown the top 2 most commonly confused terms in AI — 1) ‘Machine Learning’ or ‘ML’ and 2) ‘Deep Learning’ or ‘DL’ and the differences between them. Ready? Here it goes –
- ‘DL’ is a subcomponent of ‘ML’.
- ‘ML’ is a subcomponent of ‘AI’.
- The end.
Too simple? It got us too when we first wrapped our heads around it. Here is a visual representation to ensure that you are with us.
Okay, now that we have overcome that challenge, let’s dive right in.
‘Deep Learning’ is a highly specific technique used for realizing and enabling ‘Machine Learning’. It is the direction that computer scientists are taking towards mimicking how the human brain understands and processes information. ‘Deep Learning’ requires tons of data and extremely long periods of training in order to achieve the level of accuracy needed before it can be put to use. Of course, the problems that are expected to be solved by ‘Deep Learning’ techniques are the ones that are more complex for the broader description — ‘Machine Learning’. The challenges faced by self-driving car manufacturers is a classic example of the need for Deep Learning. Let us explain.
Self-driving cars have become real intelligent real fast but the issues they face now are not related to what the cars can do, but what the cars yet cannot do. Ex: Self-Driving cars have become more accurate than humans in identifying objects around them such as other vehicles and pedestrians, but they are unable to determine whether the pedestrian is about to cross the road or not. Now why is something so intelligent not able to solve this problem? Here is why — humans can understand another human’s posture, facial expression, age, culture, their level of readiness on the road and decide whether to keep driving, slow down or come to a complete halt. Self-driving cars are having challenges in being able to read human posture, facial expression and other subtle human signals. This is where ‘Deep Learning’ comes in. Self-driving cars will need to be programmed with an enormous amount of pedestrian body language and facial expressions and the likes before it can accurately compute how it should best respond to situations.
‘Machine Learning’ is the broader spectrum under which problems that are less complex can be solved by using existing algorithms or fine tuning them. Sure, it needs data input and goes through an iterative process, but these are problems that have a lot less complexity than the problems that require ‘Deep Learning’. The suggestions you get on YouTube, Netflix, Social media feeds and shopping platforms like Amazon are great examples of Machine Learning. Here is an over simplified version of how such systems work:
‘Artificial Intelligence’ is the broader spectrum under which ‘Machine Learning’ and ‘Deep Learning’ reside.
The concept of AI has been around for more than 200 years.
Going back even further, the Great philosopher Aristotle laid the foundation for AI through his famous work on the theory of ’Cause and Effect’. ‘AI’ first reached mass popularity in 1939 when ‘Westinghouse Electric Corporation’ in Ohio displayed the robot named ‘Elektro’.
‘Elektro’ responded to voice commands, identified colors, blew balloons and even, smoked! (like that was really necessary). Since ‘Elektro’, ‘AI’ concepts have gained more and more investment and backing to the point where we are carrying ‘Siri’ and ‘Bixby’ in our pockets today.
Now let’s reconnect on what ‘AI’ means to the world of Healthcare.
A breakthrough by the ‘Machine Learning’ team at ‘Stanford’ and ‘Intermountain Healthcare’ in September 2019 resulted in the birth of ‘CheXpert’. An ‘ML’ enabled chest X-ray interpretation program that can diagnose pneumonia more accurately than the average radiologist! This has created the argument about the need for radiologists in the future. The truth behind whether we will need radiologists or not, only time will tell but this is a strong (really strong) discovery that everyone in the radiology business need to follow (even be ready to embrace) closely. This is a minor example of how ‘AI’ is changing the world of Healthcare.
IBM Watson is playing a key role in enabling ‘AI’ studies by providing super computers that help compute information that would take years, into days and thus thrusting the capabilities of scientists to invent solutions faster and reduce expensive trial and error. ‘Atomwise’ and ‘Enlitic’ are great examples of companies using ‘ML’ and ‘DL’ to detect cancer at very early stages and use super computers to narrow down on which mix of compounds would make the best drug to fight those cancers.
One of our favorites is ‘Human Longevity Inc.’ This company is using genome sequencing to revolutionize ‘precision medicine’ by going to the level of predicting the possibility of a disease before it has actually happened. One of their approach to healthcare (taking cancer as an example) is to identify which genomes in the body are likely to mutate and develop specific drugs or medical management plans to delay or avoid the mutation from happening.
These are serious breakthroughs in the world of Healthcare enabled by ‘AI’.
We at ‘Cloud Solutions International’ closely follow global technology trends that are very specific to healthcare. We do this through our team of industry research experts and world leading market research partners. We invest in such research on behalf of our clients. With over 150+ domain experts in the field of technology and healthcare, we stay on top of where Healthcare technology is propelling towards and we stay in harmony with such advancements by aligning our solutions.
We believe that being ahead and in harmony with global trends and adapting really fast is fundamental to producing thriving, sustainable, solutions to our Clients.