PHD Discussions Logo

Ask, Learn and Accelerate in your PhD Research

Question Icon Post Your Answer

Question Icon

2 years ago in Base Papers By Raghav V

Should I include older, classic base papers from the 1990s in my literature review for a fast-moving field like deep learning?

I'm compiling base papers for my deep learning PhD. Some early backpropagation or CNN papers from the 80s/90s are cited as foundational. But the field has changed so much. Is it necessary or even useful to include these, or should I focus on the last 5-10 years?

All Answers (1 Answers In All)

By Jahnvi Answered 1 year ago

Absolutely include them, but strategically. I recommend a "historical anchor" approach. Select 2-3 true origin papers (e.g., the 1986 Rumelhart et al. backpropagation paper, or LeNet-5) not to detail their obsolete technical specifics, but to frame your review. Briefly explain their core conceptual breakthrough and how it enabled the modern field. This demonstrates you understand the lineage of ideas, not just the latest trends, which adds tremendous scholarly depth. Then, pivot quickly to show how those concepts were scaled with data and compute (the 2012 AlexNet paper is a key pivot point). This shows evolution and justifies your modern research gap.

 

Your Answer