Case Study

Should a doctoral researcher apply baseline-first advice meant for product teams?

Case context: A doctoral researcher is building a new ranking algorithm for a conference paper. A colleague recommends starting with a simple working version, then studying mistakes and improving it step by step.

Question: Should the researcher follow this baseline-first advice exactly as given? Why or why not?

Sample answer: Not exactly. That guidance is designed for people building AI products or applications, where a quick baseline and error-driven iteration are especially useful. Here, the main objective is a research contribution that is novel enough to publish, so the advice does not fit the project’s primary goal in the same way. The main point is that product-development guidance and research-planning guidance serve different purposes, and research-specific advice is handled separately.

Key points:

  • Advice is aimed at product or application settings
  • The project goal is publication-oriented research
  • Product iteration and research planning are not identical
  • Research guidance is treated separately
  • The answer should note the audience mismatch

Rubric: Full credit: correctly identifies that the advice is intended for application builders, explains that the student’s goal is research publication, and concludes that the advice is not the best direct fit while noting that research is addressed separately. Partial credit: identifies the mismatch without a complete explanation.

0

1

Updated 2026-08-12

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Machine Learning Strategy

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Yearning @ DeepLearning.AI