[D] [P] Some ideas and speculations on the problem of efficiently processing length-irregular data

Hi folks. I'm a new DL researcher who just began working at a startup that focuses on AI-based drug discovery. I'm afraid this is not the most suitable place to post this since this is more of an engineering idea, but I wanted to hear what you guys think about it and if you have any idea.

I don't know how many of you have encountered the same efficiency issue before, but I've repeatedly come across this theme while implementing …

data ideas machinelearning processing