College of Computing must make interdisciplinary work its prime directive
Computing is an important tool for problems that ultimately belong to other fields
The launch of the $1 billion Schwarzman College of Computing will mark the most significant change to the structure of MIT since the 1950s, and is already making waves in the worlds of tech and higher education. While the creation of the new college has the potential to make a significant positive impact, its success hinges on its ability to place its interdisciplinary vision, rather than aspirations of computational pioneering, at the center of its implementation. The vast bulk of the world’s most important problems and most interesting questions exist in the physical realm rather than the digital one, and while computation does well as a standalone field, it shines best when used as a tool for facing these problems.
The College of Computing is intended to “integrat[e] AI studies and research with disciplines throughout MIT to a degree and with an intensity that, it is believed, is unmatched anywhere else,” according to MIT News’s FAQ. Its announcement comes at a time when computing is becoming ubiquitous. Every discipline has unique needs with respect to computing — for instance, urban planners must now consider autonomous vehicles when designing cities, economists are increasingly integrating machine learning into their data analysis practices, and neuroscientists are deciphering the pathogenesis of diseases and complex disorders with transcriptomics. MIT needs to consider precisely how computing and AI will be integrated with these different disciplines. How will the 25 bridge faculty of the college work? And if the college is truly cross-cutting, how exactly will it interact with all the other schools aside from just these 25 faculty members? The Institute must not neglect the details of how possible research collaborations might work as it continues to plan the college.
One such detail, for example, is the building that will house the college. The architecture of MIT’s Main Group promotes cross-disciplinary work by physically placing faculty of different fields near each other. Researchers can often walk a short way down the hall and encounter scholars from other fields. If the College of Computing is supposed to interact with all departments, perhaps its faculty — especially faculty who are already considered as bridges between departments — ought to be distributed among existing buildings.
Furthermore, MIT should pay close attention to the college’s role in education at the Institute. The College of Computing has the opportunity to serve as a model for how to integrate computer science education into other fields, especially in the undergraduate curriculum. It will be uniquely positioned to provide students with the knowledge and skillsets necessary to combine distinct skills. But how can the college create students that are fluent in both computing and their primary field without simply absorbing those students into computing, with their primary field as an afterthought? Students who currently want to gain experience in computational thinking must take classes in the Electrical Engineering and Computer Science department. Unsurprisingly, this encourages students to simply declare majors or minors in Course 6, and it also misses opportunities to tailor computing classes more appropriately to each field.
Relatedly, a concerning aspect of the dialogue surrounding the new college is the emphasis being placed on student demand for more computational thinking classes, highlighted for example in an August MIT Technology Review article by President Reif which markedly portends the college’s announcement. While it is vital that MIT keep students heavily involved in every decision that affects them (not limited to the College of Computing), this particular case has one important issue. MIT students, while often aspiring to make a genuinely positive impact on the world, are primarily looking for employability and financial security from an education, and are perfectly justified in doing so. A computation-focused degree is one of the best ways of achieving these goals, regardless of the relative good it does for the world compared to other career pathways. MIT should be mindful of this when basing its decisions regarding the best course of action for humanity as a whole on the choices students make. It must stay true to its mission of increasing the connectedness of computing and other disciplines even if students continue to enroll in Course 6 en masse, by tailoring its computational engineering offerings more heavily towards their uses in the physical world.
One example of a department that would benefit from tailoring of computation classes is Biological Engineering. Currently, BE undergraduates have very few classes to choose from if they would like to focus their studies on “computational and systems biology,” an optional concentration of study that BE undergraduates can choose to pursue. Moreover, many classes in the concentration, such as Introduction to Machine Learning (6.036), belong to other departments. BE students would benefit from a machine learning course that gives them the tools and experience to address questions relevant to drug design optimization, image processing, protein interaction modeling, and a myriad of other highly contextualized problems in sub-fields of biological engineering.
A class where contextualization has been implemented well is thermodynamics. From Unified Engineering: Thermodynamics (16.004) for Aerospace Engineering majors to Thermodynamics and Kinetics (5.60) for Chemistry majors, the many flavors of thermodynamics allow students to obtain skills relevant to their future career goals. A mechanical engineer would not benefit as much as a biochemist from learning about the thermodynamics of enzyme reactions. MIT should consider models like thermodynamics as it plans how to integrate computer science education throughout the undergraduate curriculum. Computing is ultimately a tool for many disciplines, and it should be treated as such, not just as an end in itself.
Lastly, the college has the opportunity to both study and teach students how computers fit into every aspect of the modern world, from how the military uses of data science to gather intelligence to how the changing privacy landscape will affect people of the future to what AI will mean for the economy. How are computers changing our lives, and how do we want them to change our lives? These are questions we should consider as we continue to integrate computing more fully both into MIT and into society, and they are questions that, again, the college will be uniquely positioned to answer.
If the college fails in its mission, MIT lurches dangerously closer to becoming an “Institute of Computing.” Vast funding and resources will be devoted solely to innovation for innovation’s sake, undergraduate enrollments in every other field will dwindle, and, worst of all, nobody will even know what it all means. With careful deliberation and decisive action, the Institute can avoid these worst-case scenarios.