AI Without Math

AI Without Math

Making AI and ML comprehensible

Top 11 questions to ask your AI developers.

  1. What is the source and the range of the data used for training the model?
  2. What is the size of the dataset? Is the data adequate to make the prediction?
  3. What procedures have been done to clean and transform the data?
  4. Was there a lot of missing data? How did you deal with it?
  5. Are you following the data related regulations?
  6. Is there proper validation done on the algorithms?
  7. What is the range of the predicted values, and the related accuracy level and error distribution attached to the prediction (what is the tail risk?)?
  8. What kind of validation and testing procedures did you perform?
  9. How would the predicted values change as new data comes in?
  10. What other algorithms/models can be used for the prediction purposes?
  11. What other risks we should be aware of from deploying this model into practice?