Reading a Systems or Machine-Learning Paper
A paper is a claim, a method and evidence, in that order of importance. Reading the evaluation before the method shows what is actually being demonstrated, and the related work shows what the authors compared against and what they left out.
What a learner can do afterwards
- State a paper's central claim in one sentence and name the evidence offered for it
- Find the assumption the result rests on and say what changes if it fails
- Name a baseline the paper should have compared against and did not
1 · Read
A paper is a claim, a method, and evidence, in that order of importance. Say the central claim in one plain sentence, then point at the evidence offered: tables, graphs, experiments. A chart where a model wins on one easy data set supports only that set, never the broad claim of being better in general.
Read the evaluation before the method to see what is actually demonstrated. Early retrieval tests fixed a set of documents, questions, and relevance lists, then scored recall, the share of relevant items found, against precision, the share of returned items that were relevant. Returning more lifts recall but usually sinks precision, so a paper quoting one side owes you the other.
A number means nothing alone: scores need a baseline on the same setup. Ask what the authors compared against and which strong baseline they left out. Authors write to be believed, so introductions frame the gap, related work positions the rivals, and evaluations spotlight the wins; read like the editor, not the audience.
Hunt the load-bearing assumption the result rests on, and say what changes if it fails. Then check the tools of persuasion: hard numbers push belief, vivid stories push care, and both can hide what was never shown.
Claim in one sentence, evaluation before method, baseline beside every number, assumption under every result.
2 · Watch
Take it off screen
Where it sits
8 questions wait behind this lesson, each with its answer explained. Every answer feeds the sky: stars light as they are learned, and dim when it is time to come back.