Alumnus Mauricio Santillana '08, tracks disease and improves intensive care environments at Harvard Med

Mauricio Santillana, Ph.D. CSEM ‘08 Assistant Professor, Harvard Medical School Faculty Member, Boston Children's Hospital Computational Health Informatics Program Associate, Harvard Institute for Applied and Computational Sciences Former Supervising Professor: Clint Dawson

What inspired you to a career in computational science?

As a physicist, I was trained to identify, study, and characterize the patterns and mechanisms driving natural phenomena. This approach helps us understand how things work and make predictions of yet unobserved phenomena based on our prior observations. We used mathematics to synthesize what we saw in this constantly evolving world. During this time, it became clear to me that my childhood interest in programming hand-held calculators and personal computers to solve simple problems and create games (before iPads existed!) could enhance my professional life as a scientist. I realized that using computers intelligently could help decision-makers assess the effectiveness of strategies aimed at solving societally relevant challenges.

Describe computational science to a non-expert.

While extremely useful, analytical methods in mathematics using only our minds, pen, and paper, have limitations. For example, when attempting to solve challenging real-life problems, one frequently needs to compute the solution of massive amounts of small and simple mathematical problems in order to have a better understanding of the original problem. This is challenging for our minds, since it involves memorizing the outcomes of many smaller mathematical problems and then combining these outcomes to design a solution. Think about modeling the traffic in a city to assess the impact of constructing a new highway. In principle, all we need to do is simulate how a person goes from point A (their home) to point B (their office or their children’s school) and back daily. Simulating one trip may be simple but simulating how drivers will interact with one another so that each person can get from their point A to their point B and back is not easy to do in our minds or using pen and paper. While humans have a limited capacity to do this, computers can be programmed to do this all-day, tirelessly. This is the basis of computational sciences.

Describe your current job or research.

In my research, I use machine learning and scientific computing to extract patterns and signals from multiple data sources to predict the likelihood of an event given a set of conditions. I do this by analyzing our previous experiences and historical measurements. For example, my team and I calculate the likelihood that an epidemic outbreak is happening now (or will happen a few weeks from now), in a given region of the planet, given that (a) multiple people are searching the Internet for information about a specific disease, (b) doctors are prescribing specific medications, (c) weather patterns seem similar to years when epidemic outbreaks have happened, (d) people from a region with recorded cases of a certain disease are traveling to the location of interest, and (e) the quantity of disease-transmitting vectors (think mosquitos transmitting Malaria or Dengue) is abundant. I also study the likelihood that a patient may experience a desired or undesired outcome in intensive care environments.

What’s the most challenging part about your work?

The two most challenging aspects of my job are: (1) to extract meaningful signals from noisy data streams and (b) to explain to decision-makers how our computational models work so that they can trust the solutions (or likely scenarios) we present to them.

How does computational science “change the world?”

Our daily activities and the pervasive presence of sensors (mobile phones, city cameras, satellite images, weather stations, our intelligent appliances and smart homes) are generating incredible amounts of data that, if used responsibly, can help us better understand how we are changing our planet and how to best do this in a healthy and sustainable way.

Where do you see the field in ten years?

Computational sciences will be the backbone of the modeling and simulation systems that will help us better understand the vast amounts of information we constantly collect with sensors placed on individual humans, and in our homes, streets, means of transportation, environment, etc.

What is a memorable experience at the Oden Institute?

During my years at the Institute, I enjoyed learning side-by-side with very talented colleagues and mentors. Our research advisors reminded us that when computational sciences were first used to inform how bridges would be built or how wings of airplanes should be shaped; the way they would convince people of the rigor and trust they had in their calculations was by offering to be the first ones to cross the resulting bridge, or to be in the first flight on the plane designed with their results. This commitment to delivering reliable computations to transform the way we do things in the world is the philosophy taught and practiced at ICES.