Research Question
A research question states what unknown a run, paper review, or analysis is meant to reduce. A vague topic is not yet a research question.
Weak:
Stronger:
Good Questions
- Name the task and metric.
- Name the data split or benchmark.
- Name the baseline or prior belief.
- State the expected mechanism.
- State what would change the next decision.
Checks
- Can the question be answered by a concrete observation?
- Is the scope small enough for one experiment or paper note?
- Does it avoid bundling multiple questions?
- Is the question public-safe, without internal project names or unpublished claims?