09/01/2026 10:31 AM
09/01/2026 10:31 AM
PACES Model:
Problem
Approach
Contribution
Evaluation
Substaniation
How does Evaluation and Substantiation differ?
P & A: Get a 3 sentence summary.
First pass: Read every first sentence and skip all equations (also for me also look at figures)
Contribution: Is there a new problem? Or a new approach to an existing problem? Look at the end of introduction to see how they say it, but you should be able to state it in your own way by reading the rest of the paper.
Evaluation: For a theoretical paper, is there a proof? For an empirical paper (most papers in ML lol), what benchmarks, what metrics? Are there holes in the benchmark?
Substantiation: This is the opinion part, based on the contribution and evaluation, do you think the evalatuion is sufficient to substnatiate? Is the contribution enough for novelty?
Book recommendation: Make it stick (the science of successful learning)
One recommendataion is to actively make guesses about what you will see later. Or in this case after your first pass make a mental note about what you think the PACES is.
Don't spend more than 60 minutes on the PACES per week.