Reading With Purpose
Reading papers is a basic skill in research. We read papers to learn new material, identify constructs that we might need in our work, review submissions for conferences/journals, or for class presentations. And yet most training in reading papers is on the job training that you figure out as you go along.
“Training in reading?”, you say. “What, are we five?”.
Well no, but reading for research is a specialized skill that isn’t the same as reading a novel or any kind of lay reading. It requires critical thinking, the ability to skim the surface till you find a place to dive in, an understanding of how a paper is built, and the skills to abstract out (quickly) the main essence of a work without getting bogged down in details. And every use-case for reading a paper requires a different combination of these skills.
Hence the title: We will learn, via in-class activities and guidance, how to read a paper with a particular purpose in mind. Since different areas of computer science and related disciplines have very different paper styles, we can’t really cover all of them. So I’ll focus on what I know: theoretical papers with possible side excursions to papers in machine learning/data mining that combine detailed mathematics with experimental support.
Some of the skills you learn here should be generalizable to other areas. Some will be more math-specific. You have been warned