How to start research in code translation
Hi everyone,
I’m an undergrad majoring in Computer Science and this is my first time getting into research. I’d really appreciate some advice.
My supervisor suggested that I explore the area of code translation (translating code between programming languages). I’ve read a few papers and have a basic understanding, but I’m still struggling to connect the dots because the field feels very broad — ranging from rule-based approaches to traditional ML and now LLMs, as well as different levels like function-level vs repo-level translation.
I also don’t have much experience benchmarking AI models yet (I’ve mainly worked on agent systems before). It seems like performance without AI/LLMs is quite limited, so I’m wondering if getting up to speed in this area would take a lot of time.
I have a few questions:
- How should I narrow down the scope when starting in this field?
- Are there any recommended directions or subtopics for beginners?
- How do you usually extract research ideas from reading papers?
For now, I plan to continue reading more papers, but I’d really appreciate any guidance or suggestions from people with experience in this area.