When Laurel Springs School senior and Bay Area, California, native Rohan Adwankar got his AP Computer Science Principles exam back, he wasn’t hoping for a flawless score. After all, he hadn’t grueled over his studies for days or weeks.

And yet, when his marks came back, he discovered he had not missed a single question. If it wasn’t an unbreakable streak of perfectionism and countless nights sacrificed to studying that earned Rohan this impressive grade, what was it?

“I was curious about computer science,” the senior puts simply. “Not just for a test, but I was curious about the subject. Like genuinely curious.”

That’s right — Rohan is the name and computer science is his game. He says he takes to the subject naturally, but don’t get him wrong. The AP exam was far from easy, as his time spent studying and taking practice test after practice test shows. Still, he credits his success to an inquisitiveness that runs in his blood — and being a methodical researcher and somewhat of a perfectionist helps.

More than a movie for Rohan Adwankar

Have you ever watched one of those cheesy scenes in a movie during which the software engineer/computer scientist/hacker is furiously hammering away on a keyboard while blocks of green text fly by on the screen? All while the nervous cast crowds around behind the desk, waiting for the programmer to crow that “he’s in!”?

Does it come as any surprise that those theatrics are, well, just that — theatrics?

In the real world, most computer scientists use their knowledge for more…benevolent causes. In fact, in Laurel Springs’ AP Computer Science Principles class, students study the potential social uses of computer science. Rohan, for example, plans to dedicate his service to the area where the public health and computer science sectors overlap.

Turning computers into students

The concept, Rohan explains, is called machine learning. Put simply, this concept works by teaching a computer through algorithms to make discernments that normally only humans can make — for example, feeding the computer images of dogs and cats until it is able to predict which of the two animals is in an unlabeled picture that it’s presented; or showing the computer images of people and telling the computer that the smiling people are happy and that crying people are sad.

Eventually, with enough data, the computer, after being fed an image of a crying person, can output that the person is sad. As simple of a concept as it seems for humans, it is as the name suggests: scientists take machines, and they help them learn, similarly to how human babies learn.

One time, Rohan was reading Ralph Waldo Ellison’s “The Invisible Man” and decided to experiment with the book. He wrote a code that reads the book and uses a linguistics-based type of machine learning, called natural language processing, to determine the narrator’s sentiment or tone based on a given amount of text sampled at any point in the book.

What he found is that the computer could, in fact, discern between a positive and negative tone in the sample. Rohan has applied machine learning in multiple instances and discovered how closely machine learning and literature are intertwined — how much computers can learn about human behaviors just from the way we write and talk, and eventually predict those behaviors.

Then, at the Regeneron International Science and Engineering Fair, Rohan taught a computer to utilize algorithms to simulate new potential NRTIs, which are drugs used to treat HIV. In essence, the computer is taught the molecular structure of the drug in question and what it is meant to do to the human body. The computer then simulates an uncountable amount of molecular combinations until it finds a combination that is predicted to have a similar result as the already-known drug.

“The computer was taught to think about and identify which potential molecules could be successful as anti-HIV drugs,” Rohan says. “This shows how computer science is used in a broader context.”

The idea is still largely experimental and imperfect, especially when it comes to pharmaceuticals. The new generation of machine-powered medicinal discovery is not quite on our doorsteps, but the future implications are significant, especially when we leverage it against an urgent — but often overlooked — public health crisis.

The social implications of computer science

The prognosis for most Alzheimer’s patients is bleak. According to data from the Alzheimer’s Association, most people are expected to live another four to eight years after diagnosis. Early detection and preventative care are paramount to improving the outlook for dementia patients, and work done by computer scientists like Rohan helps to brighten that outlook bit by bit.

With machine learning, computers are able to detect patterns so slight that they go unnoticed by human observers. Rohan worked on a program in the Laurel Springs AP Computer Science class that utilizes the subtle nuances of human speech to detect the early presence of Alzheimer’s evidence.

“Using linguistics and computer science is how the Alzheimer’s detection algorithm works,” Rohan says, “because those computational linguistics help the machine learning model pick up on patterns in the way people with Alzheimer’s might phrase their answers to a test’s questions.”

More specifically, Rohan researched how demographic biases — how people’s speech habits and patterns based on their socioeconomic factors might alter their answers — can impact how accurate the computer is. The systems may be imperfect and take a lot of teaching to learn, but the eventual impact on public health is invaluable.

The future of computer science is strong

Rohan doesn’t have a specific college picked out yet, but he looks forward to continuing the study of how medicine and computer science can work together. He credits Laurel Springs for the integral role they have played in setting him up for a successful future.

“My experience at Laurel Springs was profoundly empowering because it allowed me to take the AP courses I wanted with the increased flexibility that comes from being able to set my own pace. This excellent system combined with the supportive and insightful faculty allowed me to truly learn the material as well as giving me the freedom to successfully apply them in extracurriculars competitions and testing.”

His advice to younger students is to stay methodical and gain a strong set of coding and problem solving fundamentals.

“When you’re coding, you need to compartmentalize problems,” he explains. “Learn how to research solutions and overcome obstacles. Also, stay up to date with different software and new information that comes up. Different models or methodologies may be more effective.”

Share on social media