Hi guys, i have my graduation project for this last year in university(AI major). It’s been three weeks but I have not settled on an idea yet(looked into LLMs and AI agents). I really want something that’s very interesting and creative(not another medical diagnosis system, and please not another sentiment analysis project) yet addresses an actual problem. Something that unexpectedly merges many feilds, or something that’s just really new. If you have any suggestions or ideas, I would be eternally grateful.
1 Like
I don’t have an answer to your question. It seems like you don’t prefer a few project areas. It would be helpful to the reader if you suggested a few areas of interest as well.
On a somewhat related note, how about using an LLM (keeping temperature in mind) to generate project ideas?
I am interested in astronomy, i did try using LLMs and adjusting the temperature to generate more creative ideas, but nothing really landed.
Have you seen this link?
If that doesn’t work out, try your luck on a freelancing site.
Hi, Konouz_Abdelaziz!
If you are interested in astronomy, you are probably well aware that machine learning and deep learning have made serious inroads into the subject, so I will not give you any tips in that direction. However, thinking about your original question, I am always piqued by the statements that AI is going to solve all of our problems, including climate change. This is sometimes even used as an excuse for the energy-hungry data centres “we” use. Personally, I would like to see whether these claims are founded. Can machine learning help there? Can we go back, look at the various revolutions (industry, Internet, whatever…), verify exactly how they affected humanity, and based on that predict on how AI is going to “save” us? Alternatively, you can try and build one of those solutions that are going to save humanity (or, even better, the vast majority of current living species) from the climate catastrophe. Your result might then be an indication that there is some truth in the earlier mentioned claims.
Cheers,
Dave