News
Literature Research and API Usage for Course Project
Written on 26.12.2025 21:36 by Ingmar Weber
Dear all,
Thanks again for your project pitches last week.
Thanks to Annika, we have a dedicated MS Teams chat for each project. Let us know if you have questions specific to your project. The more specific your question the higher the chance for a timely response.
For doing literature research, the first place to start is usually https://scholar.google.com/. Among other things, this also shows you other work that has cited the work you're interested in. Often, later follow-up research is also related. But one big drawback of Google Scholar is that you have to know the right terminology. Especially for interdisciplinary research this can often be difficult.
So I'd also definitely suggest to supplement searches on Google Scholar with searches on Gemini, ChatGPT or similar tools. Here, you can more flexibly describe what you're interested in. For best results, use their "deep research" modes (which will take several minutes). As usual with these tools, you will have to double-check everything they report. So, if we later find a non-existing reference as you blindly copied something from ChatGPT, then you have been warned. (Still, especially in deep research mode the problem of hallucinated references has largely, but not fully disappeared.)
Many of you will also want to use LLMs for annotation. In case you're access to resources at UdS or elsewhere don't cover the needs for the course project please get in touch. We can cover a small amount (up to ~EUR 50 per project) for API usage or other project-related costs. For this the easiest is if you use the OpenAI API. In this case, I'd just add to as team members to custom projects with a fixed budget (https://platform.openai.com/settings/organization/people/members). In case you really, really need a different API, then https://openrouter.ai/ is the best option. For this, please get in touch before using this. (Unfortunately, here we'd most likely have to reimburse you as we don't have a "team" set up here. So we'd have to first guarantee that you don't accidentally exceed any agreed upon quota.)
I'll be back in the office on Monday, January 8, 2026. Until then my response time will be quite slow.
Enjoy your holidays and all the best for 2026!
Ingmar