Hi all, good to be back with you.
I know randomness is a crucial part of some data modelling techniques and stuff like Random Forest, but what are ways that MMI could enhance AI or machine learning or vice versa? I find it interesting that chatgpt is this huge viral phenomenon and part of its charm is its non deterministic nature.
Is it possible to use machine learning to increase bias amplification?
Can you use MMI to enhance AI software?
AI can’t boost bias amplification, such as it is, because my bias amplification algorithms are nearly 100% efficient. However, I have tried several machine learning or AI models to boost effect size in MMI. I found that the AI can approximately double the effect size versus the simple frequentist methods commonly used.
Absolutely, MMI can enhance the production of more accurate and faster results, but both the MMI and the AI side would have to be designed to work together to produce the desired/intended outcome. This would be a significant development effort, and no one is ready for that… yet.