Open acphile opened 1 year ago
Hi, thanks for asking these great questions!
[1] https://arxiv.org/pdf/1703.04730.pdf [2] https://arxiv.org/pdf/1905.13289.pdf
Thanks for your answer. Regarding the score issue, to what extent can we leverage the numberical scores? For example, could score |x|<eps (close to zero) be viewed as non-influential data point?
It is task-dependent, I guess. For example, consider if you arbitrarily scale the loss by a constant; would that change the influence values? (I actually don't know.)
Practically speaking, how about plotting things on a histogram and observing which ones are outliers?
Hi, thanks for your nice work. I am planning to use influence functions in other settings and I have two question: