Ben Höltgen
@benhoeltgen.bsky.social
Machine Learning in Society | Philosophy of Science | PhD student @ Uni Tübingen | fm.ls/ben
Come talk to me at EAAMO in Pittsburgh next week!
October 30, 2025 at 5:44 PM
Come talk to me at EAAMO in Pittsburgh next week!
...Fairness Through Unawareness (omitting group attributes) can significantly reduce outcome inequality!
...it is often possible to reduce outcome inequality without reducing accuracy!
...Logistic Regression with group attributes is particularly prone to exacerbating inequality!
...it is often possible to reduce outcome inequality without reducing accuracy!
...Logistic Regression with group attributes is particularly prone to exacerbating inequality!
October 30, 2025 at 5:43 PM
...Fairness Through Unawareness (omitting group attributes) can significantly reduce outcome inequality!
...it is often possible to reduce outcome inequality without reducing accuracy!
...Logistic Regression with group attributes is particularly prone to exacerbating inequality!
...it is often possible to reduce outcome inequality without reducing accuracy!
...Logistic Regression with group attributes is particularly prone to exacerbating inequality!
If you assume there is a true distribution from which we can draw iid, then with enough data we can approximate it, so there can only be an epsilon-RCP because we are delta-close to the assumed true distribution?
March 26, 2025 at 5:07 PM
If you assume there is a true distribution from which we can draw iid, then with enough data we can approximate it, so there can only be an epsilon-RCP because we are delta-close to the assumed true distribution?
Is the intuition correct that, if we draw iid from a true distribution, then if two models disagree enough, at least one is far from the true model. So we can improve the models and with more data, both models approximate each other and in the limit the true distribution?
March 26, 2025 at 5:04 PM
Is the intuition correct that, if we draw iid from a true distribution, then if two models disagree enough, at least one is far from the true model. So we can improve the models and with more data, both models approximate each other and in the limit the true distribution?