[nolaspring]

Richen the dataset it’s trained on enough so that the model is correct before you release it to prod.

[IncRnd]

That's sort of obvious. How do you know that wasn't attempted?

[ironmagma]

Even if it was fixed, in a probabilistic system like this, isn't it basically guaranteed to happen with some inputs?

[MBCook]

If it was we wouldn’t expect this problem to occur, correct?

[belltaco]

That makes it sound even worse that they knowingly released it without fixing it.

[shadowgovt]

We don't actually know how to do that, or how rich is "rich enough." It's an open avenue of research to be able to extrapolate how well-tuned a neutral net is on data not in its training set.

Not to imply the problem is unsolvable, just that if an institution has zero tolerance for this mistake, the fix your describing is no guarantee it won't occur.

[6gvONxR4sf7o]

That’s not quite complete, right? It’s that we don’t know how to do that without sacrificing other things.