"My parents were great at showing me that, with whatever I spend my time doing, I should do it thoughtfully." These days, Parker Malachowsky is thinking deeply about intuitive design and machine learning in the healthcare industry. Here, he shares that whether you're in design, technology or business, striking a balance is key.
What project are you working on right now?
Currently I am working with my pal and classmate, Louis Harboe to build an interface that allows someone who has no experience with machine learning to get a sense of how a special decision-making system, called a neural network, reasons and comes to conclusions. We hope to eventually apply this work within the healthcare industry, as there are very accurate predictive networks being developed today that could be of use to clinicians, yet they can’t and won’t be used as longs as doctors cannot interpret the logic behind them.
How did you come up with the idea for your project?
The interpretability of neural networks is a hot topic, both as a means to train better models and as a way to help expert professionals (i.e. doctors, lawyers, members of the military) benefit from this type of machine learning solution without having to get a computer science degree. Most researchers are approaching this as a mathematics and computer science problem (which it certainly is, for the most part), while Louis and I are working around the premise that there is progress to be made through building exportable and intuitive interfaces that sit between a neural network and its end user. Louis and I decided to work together on our Garage Experience project at the end of last year, right when I was becoming invested in machine learning and deep learning (a subset of machine learning that typically relies on neural networks). After continually hearing about the problem of neural network interpretability, I brought it up to Louis and he, given his expertise as an interface designer, was excited to work together in approaching this problem as a design challenge.
Is there a leader or mentor who inspires you?
I currently am very interested in Demis Hassabis, the cofounder and CEO of DeepMind (now a part of Google). I’m delighted by how committed Demis is to working according to an ethical code, so that he can push the field of AI forward for the benefit of everyone.
Is there an inspirational quote you live by?
“Science is magic that works.” ― Kurt Vonnegut Jr. in “Cat's Cradle”; and
“I haven't slept for ten days, because that would be too long.” — Mitch Hedberg
Tell us a bit about your journey to the Academy. What were you looking for in an undergraduate program?
I was excited about the Academy as it seemed like the only program that was absolutely committed to letting its student explore. Even when compared to other interdisciplinary programs, I felt the Academy offered more room to try out new things and get exposed to a wide range of subjects and technology.
What advice would you give to a freshman?
If you are someone who enjoys making things, be careful of loading your schedule with mostly lecture-based classes in any semester. I did this the first semester of my sophomore year and really had a hard time with it. I found that striking a balance between different classes with different types of coursework helped me avoid feeling bogged down.
Describe one thing most people don’t know about you?
A fair amount of people knows this about me, but in my spare time I like to make music on the computer.
What is the one thing your parents taught you that you’d like to share?
My parents were great at showing me that, with whatever I spend my time doing, I should do it thoughtfully.
What are the essential things you always keep in your backpack?
For better or for worse, I almost always have my laptop with me. Though, I prefer to take notes and work out ideas on paper, so I always try to have a pen and (brightly colored) composition book on me as well.
Favorite app or website?
I think it’s pretty boring — but probably Soundcloud. To sound cooler, I’ll say distill.pub.
Favorite class at USC and why?
BAEP 455: Founder's Dilemmas — Professor Noam Wasserman is simply brilliant and makes every moment of class worthwhile. The course demands a lot of self-reflection and personal growth, so by the end of it you feel better prepared to take on any challenge—not just those related to founding a business.
Favorite way to de-stress:
Music! Dancing! (Maybe running if I’m really on my game)
(Not too guilty about either but) early 2000s pop music or crossword puzzles.