Q&A: Chris Rackauckas on the equations at the heart of just about everything | MIT News
Some people pass the time with hobbies like crosswords or sudoku. When Chris Rackauckas has a free moment, he often uses it to answer questions about numerical differential equations that people have asked online. Rackauckas — previously a professor of applied mathematics at MIT, now a research affiliate of MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and co-principal investigator of MIT Julia Lab — has already posted thousands of these answers, and if you have a question, chances are he has already addressed it. Unsurprisingly, his research focuses on differential equations and computational methods – using AI and other techniques – to solve them quickly and efficiently.
During his graduate studies in mathematics at the University of California, Irvine, which earned him a doctorate in 2018, Rackauckas focused on the medical and pharmacological applications of his work. In fact, he developed the core software and techniques for Pumas-AI — a Baltimore-based company that provides software for pharmaceutical modeling and simulation — while still a graduate student. He now holds the position of director of scientific research for the company.
Since arriving at MIT in 2019, Rackauckas has found a much wider range of applications for his “accelerated” differential equation solvers, including global climate modeling and heating, ventilation, and air conditioning (HVAC) systems. ) buildings. He has put time into his efforts to find ever faster ways to attack differential equations to talk about this work, which has earned him numerous accolades, including the Artificial Intelligence Accelerator Scientific Excellence Award. of the United States Air Force 2020.
Q: How did you come to what you do today?
A: As an undergraduate math major at Oberlin College, I mainly focused on “methods courses” in science areas – statistical methods in psychology, time series econometrics, computer modeling in physics, etc. I didn’t have a well thought out game plan. I just wanted to understand how science is really done and how we know when our scientific approaches give us a correct model of a given system. Fortuitously, this path turned out to be good for someone in my current line of work.
In graduate school, I went into biology – specifically combining differential equation solvers with systems biology. The goal was to create predictive models of how the randomness of a chemical and its concentration change in the body, although at the time I was working with zebrafish. It turns out that systems biology is very close to systems pharmacology. You’re essentially replacing fish with humans.
Q: Why are differential equations so important in the world around us?
A: The way I like to describe it is that all science experiments measure how something changes. How do I go from understanding how things are changing to predicting what will happen? This is what the process of solving a differential equation is for. Simulations, which are experiments we perform on computers, can involve solving thousands and thousands of differential equations.
Such a simulation can tell you, for example, not only how the concentration of a drug changes over time, but also how the effects of the drug on the body change. It’s not the same for everyone, so you have to adapt the equations to individuals, depending on their age, weight, etc.
Q: Since you’re focusing on “speed-up” equation solvers, where can you find the best opportunities to speed things up?
A: Clinical trials of a new drug have a fixed duration; you can’t just make the human element faster. But in preclinical domain, there is always a period of analysis. It could cost $10 billion to develop a new drug, so before you start something like that, you want to know how likely a drug is to work in its target population, as well as the optimal dose for an individual. This is the goal of preclinical analysis and quantitative systems pharmacology. Say you typically spend three months in analytics and six months in clinical trials. If you can shorten this analysis from three months to one day – a speedup of about 100 times – you will have cut a drug’s release time by a third.
Then there’s clinical pharmacology, where if you can figure out how to get the first dose right, you might be able to save time on repeating trial items. It turns out that my colleagues at Pumas and I have already managed to accelerate the preclinical analyzes performed for Pfizer by 175 times. Moderna has also publicly used Pumas and our clinical testing methods in its clinical analysis of the Covid-19 vaccine and other drugs.
Here’s another opportunity to save time and money: Mitsubishi has a facility in Japan to test HVAC systems. You have to build the whole system and then test it in a building. Each experiment can cost millions of dollars. We are now working with them to test, say, 10 different ideas on a computer to decide which of these 10 options they should select for prototype and further experimentation.
Q: Can you give us other examples of the use of your work?
A: The SciML.ai website maintains a (woefully incomplete) showcase of the amazing ways people have used these methods. CliMA – an Earth system model developed by scientists at Caltech, MIT, and other institutions – relies on the differential equation solvers I wrote. Recently, I was at an applied math conference where a group, independent of me, reported that they had used my software tools to run NASA launch simulations 15,000 times faster.
Q: What are your plans for the future?
A: There are a lot of things in the pipeline. An application that I have just begun to study is forecasting the flow of forest fires; another is to predict transient cardiac events such as heart attacks, strokes, and arrhythmias. A third area I’m moving into is the field of neuropsychopharmacology – trying to predict things like individualized biosignals in bipolar disorder, depression, and schizophrenia in order to design drugs that are better suited to treating these disorders. This is an area where there is an urgent need which can lead to much more effective treatments.
Between these projects, I might take a moment to answer the odd question about differential equations. You need to relax sometimes.