Artificial Intelligence and Machine Learning Resources

First off, I found this figure very helpful in thinking about which types of AI tools I actually want to use. Using targeted AI tools, such as scite_, to help with specific stages of the research process is much more effective than just trying to get the general purpose LLMs (e.g., ChatGPT, Copilot) to do much of it at once. It helps with linking your work back to real sources, and reduces the likelihood that the AI goes on a completely unexpected tangent.

I also recommend being very careful uploading someone else's work to an AI tool. If you have not been given permission, you are sharing someone else's intellectual property with a platform that they did not approve. It may be acceptable if you understand the licensing agreement with the AI software, and know for sure that it is not copying and retaining that data for its own usage. Otherwise, it puts you in a compromised ethical situation.

Machine Learning

Cross-Validation

Here is a list of resources related to cross-validation:

• Cross-validation with SciKit Learn in Python
• Cross-validation using k-fold with scikit learn
• Python Machine Learning (ML) cross-validation
• Why every statistician should know about cross-validation
• Solving 9 common cross validation mistakes