Data + Science

‘Data never lies; people do.’

Data is an amusing artefact. It exists all around us, in many forms, and unless we know exactly how to read it, we are oblivious to it. It is no wonder that most professions are based in the ability to understand one specific type of data. We spend a quarter of our lives just learning about different types of data and figuring out which one we want to spend the rest of our lives around! What are skills if not the art of interpreting data? It is in this interpretation that lies are born from. Data doesn’t write stories, interpreters do. What comes from understanding the data is the hypotheses, observations, conclusions, drawn from the data. That is not what the data tells us; that is what the interpreter understands from the data, given their abilties.

Most of science once existed as hypotheses. These hypotheses were then proven by validating observations across different types of data points, explaining the intricacies of the data and hypothesizing exceptions. Science is beautiful because it is the truth, given a set of underlying constraints, assumptions and circumstances. The same gravity that doesn’t exist beyond our atmosphere, is a defining trait of life on earth because we can observe, calculate and explain the data that we feel and observe, given the constraint of the earth’s gravity. All of science is rooted in data but data is not associated with a science.

Given my strong affinity to data and science both, in hindsight, it shouldn’t have been surprising that I ended up working as a data scientist :).

Next Phase of Life

The last few months have been some of the most gruelling ones I have ever encountered, personally and professionally. I was supposed to graduate next month with my Masters, but I pushed that to December because a marvellous opportunity presented in the nick of time. In about 24 hours, I will leave Indianapolis, which has been my home since I arrived in USA in August, 2016, to pursue a research internship at one of the biggest technology companies in the world. Two years ago, had someone told me I’d be pursuing a research internship today, I’d have laughed really hard at that proposition. Now, I am only incredibly happy.

But the story of how I got here is very dear to me. I planned to pursue a MS, while focusing my work on embedded systems. After all, it was the ‘safe’ thing to do. It was one subject that I took in the first semester of my Master’s that made me realize that my real interest lies in artificial intelligence and the cutting edge of computer science, not embedded systems. It was way out of my comfort zone and a terrible risk. But soon thereafter, I started working on my thesis which aims at providing better preventive healthcare to patients by using natural language processing. Concurrently, I continued working on another project with another professor that involved applying artificial intelligence to the task of a utility prediction system. Both these projects that started around January, 2018, and the work I have done since then has been the most gratifying and satisfying part of my academic life. I have learnt more than I ever knew, and yearn to learn a lot more. It led me to present my work at avenues I never knew achievable. If it hadn’t been for that one choice, one leap of faith, I wouldn’t be here today, and I wanted to take a minute to absorb it and relish it, before the next phase kicks in tomorrow.

I start interning on Monday, and I can’t wait for the work phase of my life to begin. It is a big step forward for someone who has spent the last 22 (!) years of his life studying and preparing for this very moment. As of right now, I feel humbled, excited and nervous, all at the same time…

AAAI 2018

Over the last few days, I attended AAAI 2018, which is among the premier conferences on AI conducted in the whole year. I feel honored that I got a chance to present my research project under the ‘Innovative Applications of Artificial Intelligence’ (IAAI-18) track, and attend the whole conference. Having attended two much smaller and more specialized IEEE conferences in the last 6 months, the sheer scale and expert knowledge of AAAI has been an overwhelming experience.

Here are some quick observations:

• AI research is just beginning to scrape the surface. There is so much more that is possible, so much more to do and so many more things to accomplish.
• It seems pretty evident that AI is distinguishing itself as its own science field, away from computer science. Most of presentations I saw did not talk about the computational aspect, because the large time, data, capacity and computational abilities required are assumed. But so is research for most other sciences like physics, biology, chemistry, etc. AI has its fundamentals in computer science, but it is carving its own place.
• Machine Learning, Natural Language Processing, Reinforcement Learning, Computer Vision, etc. have started to branch out as distinct fields. Working in one of them does not mean you are good enough to work in any or all of them. Domain specific knowledge is valuable.
• The importance of research is extremely well understood by American and Chinese companies. I did see representation from a few other Asian and European companies as well, but what was noteworthy for me was the lack of any Indian company presenting significant research work in AI.
• Indian companies may be absent, but a lot of AI research at various institutions and universities is conducted (and presented) by Indians. I think this also exemplifies the limitations with our academic structure and the desperate need for its overhaul. Except a few professors from IITs, most of the Indians at the conference did not reside / conduct research in India.
• Research is a significant division of every Chinese technology company like Didi, Tencent, Alibaba, JD, Baidu, etc. Even after spending a couple of days in China last October, I was really surprised at the scale of the work these companies do. (Didi powers more than 20 million rides everyday!)
• Most companies are just beginning to explore and deploy the first sets of AI applications, and are already impressed. Others that have been in the area for quite a while are marching ahead at great speeds. But AI is more than just a buzzword, and it needs to be deployed with care.

I am glad I switched my research area within the first semester of starting my Master’s program. It is an exciting time to be living in, and especially so if you work in AI!

P.S.: In case you are wondering, I presented a water demand prediction model that I developed as a class project in Spring 2017, and improved during Summer 2017 at IAAI-18. While this project was in applied AI, my primary research area is applied natural language processing in the healthcare domain. Feel free to write to me if you want to know more!