Have you noticed that the recent surge of data scientists have a background in computer science? It’s not a coincidence. These two domains are important in their own right but when merged together, they produce powerful results.
We are thrilled to announce the release of episode 10 of our DataHack Radio podcast with none other than Professor Jeannette M. Wing! She has over 4 decades of experience in academia and the industry, and there is no one better to give a perspective on how computer science has evolved, and how it meshes with the data science world.
I have briefly summarized the key takeaways from this episode below. I recommend listening to the podcast to truly get a feel for how computer science and data science are a powerful combination when used together. Enjoy this episode!
Subscribe to DataHack Radio NOW and listen to this, as well as all previous episodes, on any of the below platforms:
Professor Jeannette Wing’s Background
Professor Wing has always been fascinated by mathematics and engineering since her childhood. She went to graduation school at MIT and started majoring in electrical engineering there. During her initial days at the university, she was introduced to the world of computer science and that prompted her to change majors. And there was no looking back from that point on.
Post her days at MIT (where she also successfully completed her Ph.D in computer science), she worked at the University of Southern California for a couple of years before joining Carnegie Mellon University. She was the computer science department head twice at Carnegie Mellon. In between those two stints, she worked at the National Science Foundation (NSF).
During her second time as the department head at Carnegie Mellon, Microsoft approached her and she took up a role there in 2013. Within a year of joining, she was put in charge of all the basic research labs, including in Silicon Valley, New York, Bangalore, and Beijing, among others.
And then last year came Columbia University and a chance to work in academia again. At Columbia, she is the Avanessians Director of the Data Science Institute and Professor of Computer Science. She reports directly to the President of the University.
Using Formal Methods Techniques to Improve Machine Learning Algorithms
Although there has been decades of research done in computer science to formally show how one can prove how a program is correct, this is all with respect to mathematical logic. What data science is now bringing is the complexity for proving how a property is correct with respect to inherently probabilistic and statistical methods.
Professor Wing firmly believes that a lot of the new data science methods should be revisited by the formal methods techniques. Its a challenge for the formal methods community to help data science grow using these concepts, something which hasn’t yet happened.
In case you are not aware, formal methods are mathematics based techniques especially used in computer science. You can read more about them here.
Research Projects in Academia and Microsoft
Professor Wing, in her current role at Columbia University, is working with the AI community to understand what methods and logic are required to specify the relevant properties that these machine learned models should have. She feels this will help build safe and trustworthy AI systems for the future, a topic Professor Wing is a strong advocate of.
At Microsoft, she was overlooking several research projects in multiple locations as I mentioned above. The Bangalore lab, in particular, had a couple of big strengths:
- Theoretical computer science, and
- Technology for emerging markets
Difference between Working in Academia v Industry
“I’m really just an academic at heart.” – Professor Wing
A very common question from folks new to data science is – “what’s the difference between working in academia versus getting industry experience”? And Professor Wing was kind enough to cover this topic.
She echoes the wide-held belief that being a scholar has it’s own distinct advantages. You have more freedom to explore questions like why something works, rather than just focusing on how it works (which is what happens in most industry roles). The science part of both computer and data science comes from research and academia far more than the industry.
Where are Computer Science and Data Science Heading in the Next 5 Years?
- Professor Wing feels AI, Big Data and Data Science are going to be much more pervasive in almost all sectors in the coming years
- Quantum and biological systems are areas where we should be focusing on more
- Cyber-physical systems like self-driving cars are going to an integral part of our lives
- And of course addressing problems like bias and correcting algorithms keeping the society in mind is something the research community should be focusing on
It was a privilege hosting Professor Wing on our podcast. Her explanation of formal methods and the important part they are playing in the software industry was a true delight to listen to. Fans of mathematics will surely love this episode.
You can also read this article on our Mobile APP