What I learned in a 2-week research internship

At IPTSE 2013 winter school, I did not just work on a regular machine learning project, but gathered some valuable life lessons too. My project basically involved detecting violent scenes from audio data, which my team and I were able to implement to some extent, given the large amount of data and our laptops as the only computing resources to handle it. We also kicked in Stacked Autoencoders to learn more representative features from the sound signals and hence come up with better results.

My experience at IPTSE revealed many perspectives previously hidden to me and helped me push my limits. The lessons I learnt while working on the project not only increased my technical know-how, but also taught me more than a few lifehacks.

1. Implementation is the easier part: For the most part, a research project involves designing the system by studying the theoretical background and coming up with innovative modifications and developments. Since the working model is not made for consumer purpose, one can afford to ignore many aspects like security, robustness etc. Moreover, many packages are already widely available, and only certain technical modifications need to be made during implementation.

2. You can push your limits beyond your imagination: I was in the zone for those two weeks and slept for only 5 hours on an average per day. During this time, I was working for more than 12 hours with only meal breaks in between. I didn’t even need to set an alarm to wake up early morning. Though I must admit I was more or less burnt-out at the end of this period.

3. Always be reading and building: The major part of research requires us to read lots of papers and articles and come up with our own ideas. This post explains that perfectly. However, having done that, the toughest part of the process looms upon us — taking the leap into implementation. Though literature surveys and reading up on similar research form the foundation stones, getting the gears running in your brain to set afoot into implementation, is a daunting task on its own.

4. You are not alone: Many a times I get too skeptical of myself and start questioning my knowledge and my decision to work in this domain. But when we meet like-minded people, we realize that we’re not the only ones having such thoughts. We share the same ambitions and problems (procrastination, social network addiction). This sets a common ground for us to strive together towards our goals.

And no, I did not have the craziest winters, but it was surely an experience of a lifetime and taught me things which I couldn't have learnt anywhere else. Budding researchers and hackers will find this helpful while striving towards their career goals.