daniel@blog

  • So you want to do some scientific research as a high school student?

    It is difficult to get started doing research as a high school student without any existing connections with companies or university labs. Therefore, I have compiled a list of notable activities, programs, and awards that I have heard of. Obviously, I have not included every possible opportunity, but these should...

  • The birth of a new sub-sub-field: adversarial attacks on 3D point sets

    An interesting phenomenon with neural networks is that it is incredibly easy to perturb or change the input by a little amount, and cause the network to make a completely different prediction. In general, this field of adversarial machine learning contains two main problems that everyone is slowly chipping away...

  • Coming up with new research ideas

    This post contains some notes on methods for obtaining research ideas, so I can remember to apply them when I am stuck. I believe that working efficiently is more important than blindly throwing time at a project. For my research work, I spend a lot more time thinking instead of...

  • n-grams + BK-trees: tricks for collapsing UMIs faster

    Creating a new algorithm or data structure is often not just one major insight, but a collection tricks that work together to produce better results. With that said, what kind of “better results” are we looking for? For computer scientists, this question often results in three answers: higher accuracy, faster...