We need to fix it. So with all questions, if you want to evaluate them in context, you need to be able to have a real-world feedback loop and iterate on the ground truth, and have a body and a voice for the report. So, if you take that away, now, all of a sudden, the algorithm can only learn from itself. Here comes the problem. : This is also of interest because I think both you and Shelvin mentioned appropriate or bad behavior, these are subjective types of lar and can vary greatly.
I think it's interesting that when we think Special Phone Number Data about how the field of data science is going to evolve, someone a few years ago said that this century The sexiest job will be that of a data scientist. Well, everything you two talk about suggests that the next sexiest job will be that of a philosopher.
The person being able to reason about what these algorithms should or shouldn't do is something that's really scarce for me right now. Data Science: Isn't that what Plato said years ago? : Loop back to the start? But that seems to be a rarer skill than what you call data science, this idea of determining whether something is appropriate and what those norms should be.