How to start research as a student (no experience needed)

Many students think research is only for PhD students or people in big labs. That is not true. You can start with a laptop, the internet and some time each week. Here is a simple path.

1. Pick a field you are curious about

Do not try to choose the "perfect" topic. Pick a field you enjoy, for example machine learning, data science or mathematics. Your exact question will come later.

2. Read survey papers first

A survey (or review) paper explains the main ideas and work in a field. It is the best first read. Search for "<your field> survey" on Google Scholar or arXiv. Use the three-pass method so you do not get stuck.

3. Reproduce one paper

Choose one small, recent paper that shares its code or data. Try to run it and get the same results. This teaches you the tools, and you will often find something that does not work as expected. That is often where a good research question starts.

4. Ask one small, clear question

Good beginner questions are small. For example: "Does this method still work on a different dataset?" or "What happens if I change this one step?" A small question you can finish is better than a big one you never finish.

5. Keep notes of everything

Write down what you tried, what worked and what failed, with dates. Save your code on GitHub. Good notes make writing up your work much easier, and they let others check and reuse it.

6. Write it up and share it

Write a short report: the question, what you did, what you found and the limits of your work. Share your code and data. Ask others for feedback before you call it finished.

7. Do not do it alone

Research is easier and more fun with others. Find a group of students who share your interest. You can review each other's work, split tasks and keep each other going.

This is exactly why APJ-RAISE exists: a free community where students learn, find a team and publish their work on our Open Research page.