Artificial Intelligence and Medicine: The Future is Now.

AI and the Medical Field

Aidan Ouckama

What even is considered AI?

AI, or Artificial Intelligence, is a recent breakthrough in Computer Science. Imagine Intelligence, but artificially made. That wasn't much help. AI is basically a computer that can do what a human can do, think, learn, and complete tasks that require a human-level of intelligence without the support of a human. AI is commonly described as "smart" which it technically true, but this isn't the aspect of AI that we find useful. The presicion an AI has when given data and the pin-point predictions it makes with said given data is what makes AI so useful to us humans.

Artistic representation of Artificial Intelligence

How is AI used in our world?

AI is always helpful when there are large amounts of data involved. And there are massive amounts of data in the medical field: diagnoses, prescriptions, medical documents, etc. Outside of the medical field, however, AI is used in a multitude of areas. The AI that most of you have heard of is probably video game-based AI where your enemies or allies are controlled by computers within the game. These AI give the gaming environment that they are deployed in a more realistic feel, whether it's to simulate a war or even just a peaceful village with passive NPCs (non-player characters) roaming around the terrain.

The most basic form of video game "AI". Pong was the worlds first wide-scale video game to be released and future versions of the game introduced AI.

Here's an example of a Machine Learning AI tic tac toe algorithm that I developed. It learns how to beat you as you play it! (It might take a LOT of games though... I'm still working on it)

If you're interested in looking at the source code take a look at this.

Tic Tac Toe

Other common, everyday examples of AI are Google Home, Alexa, and Siri. This software is known as conversational AI, as in they are used to conversate as if talking to an actual human, versus chatbots which are programmed with already inputted responses. This "conversational" aspect gives the AI a more dynamic feel when speaking to the home assistant, as the responses may differ and change as you tell the AI information about yourself, which the computer takes in as data to be processed in their "brain." This implements an "input to output" system described later.

How is AI used in the Medical Field?

As mentioned above, data such as diagnoses, prescriptions, and medical documents are processed by AI to do a multitude of different tasks. It crunches data and creates predictions based on the given data. An example of this is actually a topic we went over in class. AI was used to predict the protein structures we looked over when learning about the structures of proteins and how their R-groups and amino acids impact the shape. How this is done is the AI is given data about how different R-groups interact with each other and then using that data they can predict the shape of any protein when given an amino acid sequence. It is astonishing to see how many different cases an AI's input-output system can be implemented in!

An example of an AI generated protein structure. Similar to the one presented by Ms. Eckert during one of our lessons.

Another example of AI being used in the medical field, outside of the teachings of class, is the detection of cancerous / dangerous cells based off of stained tissue samples. In a recent study (2018), it was shown that researchers in Google AI Healthcare created an algorithm called LYNA (Lymph Node Assistant) that did the purpose stated above. LYNA took photos of the stained tissue as input and output areas of the photo where cancerous cells are possibly growing. They study found that LYNA was accurate 99% of the time, and also cut the average review time of doctors in half.

Samples that LYNA analyzed. In this sample LYNA found a patch of dangerous cells, bordered in the right panel.

Overall, AI is truly a really interesting subject and you don't have to be some technological genius to understand it. AI is implemented everywhere, even in places you didn't know it existed! I personally have a large facination with AI and have been recently been trying to wrap my mind around the processes it takes to created even a basic AI like the tic-tac-toe algorithm. Relating back to our class, AI was even used in our lesson plan as it was what developed the images we used to look at protein structure. The amount of overlap computer science and biology / medicine in this day and age is astounding, with medical engineering being a possible career path. The possibilities to AI is truly endless.