3 ms·
The best learning method I've experienced was in a course in which we had to summarize a relevant paper in 2 pages for the professor. I have to admit that I di
by behnamoh 5y ago
The best learning method I've experienced was in a course in which we had to summarize a relevant paper in 2 pages for the professor. I have to admit that I didn't learn anything from the professor __in__ the class, but reading a paper (which I chose) and summarizing it was immensely rewarding. In addition to reading the paper (which was pretty technical), I watched YouTube videos of the authors explaining the ideas in simple terms and read about the intuition on their DeepMind webpage. There was this comment on the paper which I read on Reddit (or maybe YT) that basically introduced me to the whole field of transfer learning, which then led me to my current research area. To think that all of this started with one paper (and another course which also touched upon transfer learning but from a causal perspective---"transportability") makes me appreciate this method of learning.
It's not even learn by doing. It's learn by not knowing. The past couple months literally felt walking in uncharted territory because no one in my school knows about my current research method (so no professor can spot my knowledge gap). I have felt down, up, excited, and anxious as I tread carefully in an area in which I have more knowledge-gap than I have knowledge. A sense of suspense and fear, but also great pleasure that comes with exploring something that few people in my field would even try. It's risky but I like it.