En-Route To Oxford: Asmita Poddar
Aditya Tripathi | Jul 29, 2019
A Computer Science engineer of the class of 2018, Asmita Poddar rose in the field of research exponentially. Apart from interning through highly competitive scholarships in countries like Taiwan and France, she has presented her research papers in multiple international conferences of high repute. After working as a Research Assistant at the National University of Singapore for almost a year, she has bagged an admit in the prestigious University of Oxford for her M.Sc. Degree. Read on to know about her exemplary journey.
MM: Firstly, congratulations on getting admitted to the University of Oxford! Tell us more about how you went through the grad school application journey and ended up in Oxford.
AP: Oxford is home to a community of world-class researchers and excellent post-graduate training. The MSc in Computer Science program at Oxford will allow me to gain breadth and depth of coursework over the first two trimesters and subsequently pursue my research interests through my dissertation during the last term and summer. Such a condensed year-long program not only provides rigorous coursework and research, preparing students for doctoral or industrial positions; but does so in a timely and economical fashion.
The collegiate system at Oxford also really appeals to me as it would facilitate great cultural and ideological exchange by meeting and living with students from diverse educational and social backgrounds. I would be a part of the Green Templeton College, a graduate college at Oxford, where I plan to partake in not only classes and tutorials, but also experience the motley of clubs and societies that define life in an Oxford College. Thus, the coming year shall broaden my perspective as both a researcher and as a complete individual.
The application process is straightforward consisting of a Statement of Purpose, recommendation letters, and TOEFL score, without any need for GRE scores. One-third of the over 500 applicants were shortlisted for an interview with an Oxford professor, of which less than 50 were eventually selected for the program.
MM: You have been at the National University of Singapore as a Research Assistant for a year. What was the application procedure for this and how has your experience been?
AP: I applied to several positions in universities in Singapore. After an online interview with Prof. Wynne Hsu, the Director of the Institute of Data Science (IDS), I was required to complete a coding task, following which I joined IDS as a Research Assistant in August 2018.
IDS and NUS, in general, provides you with excellent research atmosphere, state-of-the-art facilities and a healthy ecosystem of industry-academia collaboration. Over the course of my tenure, I got the opportunity to work on various projects in the broad domain of AI for healthcare. The first project was on estimating the uncertainty of deep neural networks for medical imaging to detect Diabetic Retinopathy. Once this was completed, I started another project where I used Natural Language Processing to perform Named Entity Recognition (problems, tests & treatments) on Electronic Health Records.
MM: You recently returned from the EEML summer school held in Bucharest, Romania. How exciting was it to attend your first international summer school?
AP: Early this summer, I received the wonderful news that I had been awarded a full travel grant to participate in the Eastern European Machine Learning Summer School. The school was impeccably organized by Romanian expats in top ML positions from around the world to strengthen and broaden the community around the eastern part of Europe. The lectures, along with comprehensive practical sessions, covered most aspects of modern ML and coupled with a Kaggle competition on AI for Social Good, made for a very exciting week of fun-filled learning. Exploring East European history and culture was also a very novel experience. I interacted with several academicians and industrial researchers from all over the world, including professors from Oxford and Berkeley, and had the opportunity to present and discuss my work during the poster session.
MM: You presented two posters at the 13th Women in Machine Learning Workshop (WiML), NeurIPS held in Montreal, Canada. Share your experiences from NeurIPS with us.
AP: I received a full travel grant for participating at WiML in NeurIPS 2018. It is one of the flagship conferences on machine learning and presenting my posters gave me a chance to showcase my work, address different questions and get new perspectives on my research directions through interactions with established researchers and graduate students from around the world. It was both humbling and exciting to chat with the likes of Ian Goodfellow about how, for example, he was motivated to invent Generative Adversarial Networks after a heated debate in a Montreal pub! The inclusion of under-represented groups was also a big aspect of NeurIPS 2018 (WiML being one of them with Black, Queer and LatinX in AI being the others).
NeurIPS has a massive industry presence with all the big & small names in tech showcasing their research and products in AI and out to recruit the best global talent. Montreal is also a charmingly beautiful city, and a walk around Old Port and a trek up Mt. Royal at sub-zero temperatures on alternatively snowy and sunny days was quite an adventure! pic
MM: Soon after your graduation, you received the coveted Charpak Research Scholarship for an internship in INRIA, France. How was your experience in France?
AP: I contacted several professors in France and was selected by Dr. Serge Iovleff from INRIA, Lille-Nord Europe after an interview. Thereafter, I applied for the Charpak Research Internship Scholarship and was among the 13 recipients of the award in India for 2018.
At INRIA, I co-created an R package for end-to-end simulation and classification of high volume, multi-dimensional and temporal spectroscopic data provided by the Sentinel-2 satellite covering the area of France for applications in ecology and cartography. Despite the obvious language barrier, INRIA has a very warm researcher community. My Professor was very affable, even inviting me over for a scrumptious BBQ spread of steak, French wine, and cheese. I also enjoyed the European culture of a strong work-life balance, traveling extensively during all the weekends throughout the lovely summer. Since it was the FIFA World Cup season, it was a lot of fun to be in Paris, when France won the Cup and erupted in celebrations!
MM: You gave an oral presentation at the Sound & Music Computing (SMC) Conference held at Limassol, Cyprus. What was your publication about?
AP: Following my work at Academia Sinica, I co-authored a paper with my collaborators in Taiwan and Austria on #nowplaying-RS: A New Benchmark Dataset for Building Context-Aware Music Recommender Systems. I learnt to publish under tight time constraints, working with my colleagues in three different time zones and incorporating reviews and comments from the two venues it was rejected before getting accepted into SMC. I flew to Cyprus for the week of July 4-7 during my internship in France to join a vibrant community at the intersection of music and computing. It was interesting to attend the concerts on computer-generated music (though I prefer the more human version for now), experiencing first-hand the results of multi-disciplinary research combining technological and artistic methodologies for modelling and generating music through computational approaches.
MM: Talking of your paper from Academia Sinica, how was your experience as an intern under the Taiwan International Graduate Program - International Internship Program (TIGP-IIP)?
AP: The Academia Sinica internship selection was a mixture of hard work, a last moment application and a very supportive supervisor who was so impressed after an interview with me that in spite of being in talks with another intern, he selected me for TIGP-IIP. Under Dr. Yi-Hsuan Yang's supervision, I developed a music recommendation engine using machine learning on listening events extracted from Twitter and Spotify user feeds. We continued to collaborate for a long time even after the end of the internship, which culminated in our co-authored paper at SMC. The TIGP-IIP is a great scheme that not only provides interns with a handsome stipend and accommodation but also lets you explore one of the most advanced nations of East Asia, Taiwan and its beautiful hot springs, delightfully green hikes, and pristine beaches while gorging on some delectable Taiwanese cuisine (though I am not a big fan of the stinky tofu) at the ubiquitous Night Markets.
MM: Your final year at NIT Rourkela seems to have been quite eventful. How difficult was it to strike a balance between placements, research work and preparations for examinations?
AP: I had taken the TOEFL in my penultimate year back in 2017 and for the GRE, I did the usual preparations from ETS, Magoosh Flashcards and the Manhattan app. Though I was quite certain that I wanted to pursue research, it helped that I had already received an offer from Centre for Development of Telematics (CDOT) by September, which allowed me to focus on publishing my paper, applying for research positions, finishing my Bachelors' project and preparing for the GRE. The final semester was, of course, a hectic one with all the farewells, photoshoots and Jashn activities; but as long as one can prioritize their goals, it's not too difficult to maintain a balance.
MM: Your research has been mostly focused on machine learning and data science. At what point in your life were you sure about your interests and chose research over a corporate job?
AP: During my second year internship at IIT, Kharagpur, I worked on a project on deep learning and was extremely fascinated by its scope. Consequently, I did multiple other projects and internships on machine learning and got further interested to delve deeper into it. Hence, I carved out a career trajectory that would enable me to pursue ML research applied to a broad spectrum of applications. With the seemingly endless possibilities and applications of Artificial Intelligence, this is an exhilarating time to be involved in this domain and going ahead, I would like to apply my research in an industrial setting, bringing ideas from the laboratory to fruition in real markets.
MM. We have discussed quite a bit about your professional and academic career, but how important have extra-curricular activities been in your life?
AP: I have always been a very active kid since my school days, playing a lot of volleyball, table tennis and badminton which continued into my days at NIT Rourkela where I was a part of the institute swimming and volleyball teams. Debating has always been a cherished pursuit, and being a part of Clarion was yet another fascinating journey as I developed skills in rigorous reasoning and argumentation while forming a close circle of lifelong friends as well as traveling to several institutes for parliamentary debates and MUNs. My time in Monday Morning lent strongly to my growth in authoring journalistic and non-technical articles in addition to developing people and interviewing skills. In my final year, I joined the Student Counseling Program in the very first group of Coordinators to lead the program from its infancy into a full-fledged platform for counseling, supporting and mentoring the students of NITR.
MM: How different is the research culture at institutes like IIT Kharagpur and NIT Rourkela compared to institutes in countries like Taiwan, France, and Singapore?
AP: The research culture in each of these countries is different, and your experience also largely depends on the team and/or the professor you are working with. My supervisors have always been very supportive and encouraging, but they all vary in their degree and nature of guidance & involvement. Of course in an institute of the stature of NUS, the available infrastructure is much more advanced and the student community a lot more diverse but it is equally remarkable that in NIT Rourkela, with our limited resources, we are competing on the world stage with our research output.
MM: How has NIT Rourkela been instrumental for your growth, both personally and professionally?
AP: NIT Rourkela has definitely been the foundation of everything I have achieved in the last half-decade. Besides the obvious benefits of being part of an institute that's ranked among the top 10 in India, I also made friends for a lifetime and built my foundations in Computer Science & Engineering. Thus, over the four years, I grew into a complete person while navigating my way through assignments, projects, exams and the meandering lanes around green hills, cool lakesides and quaint quarters of the immense campus.
MM: Finally, what advice would you give to a reader who is an aspiring researcher, what does it take to become Asmita Poddar?
AP: As a researcher, I have benefitted a lot from having a focused direction of my research goals. While it is necessary to explore different fields to find one's interests, it most definitely pays through compounding if one can find their broad domain at an early stage. It also helps to have insightful mentorship to obtain guidance in different aspects of academic and professional life.
I have always strived to maintain a balance between my intensive academic research and social & extracurricular activities and that I believe is often very important in developing a wholesome personality. Not only do you end up making friends from different walks of life but you also learn to be a very handy multi-tasker. One must also learn to have a composed, optimistic and calm self even in the most adverse of settings which empower you to face and solve challenges head-on. Along with a bit of weirdness to top it off, I guess that's all it takes to be Asmita Poddar.
Team MM wishes her all the best in all her future endeavors.