However, sampling is not the only source of uncertainty. In practice, distributions change between locations and across time.

Figure 10: Illustration of the frame-based pupil segmentation: (a) the input eye image I; (b) the generate binary mask M; and (c) the detected pupil boundary Q and the pupil center c. 16 C More Details in Experiment C.1 Evaluation metrics The detailed description of the four metrics adopted for the dataset evalution are as follows:

Unqualified images are removed as described in Appendix A.3. Was the "raw" data saved in addition to the preprocessed/cleaned/labeled data (e.g., to

Age Verification Is Reaching a Global Tipping Point. Is TikTok's Strategy a Good Compromise? TikTok's new age-detection tech seems like a better solution than automatically banning youth accounts. But experts say it still requires social platforms to surveil users more closely. Governments worldwide are moving to limit children's access to social media as lawmakers question whether platforms are capable of enforcing their own minimum age requirements.

Predictions about people, such as their expected educational achievement or their credit risk, can be performative and shape the outcome that they are designed to predict. Understanding the causal effect of predictions on the eventual outcomes is crucial for foreseeing the implications of future predictive models and selecting which models to deploy. However, this causal estimation task poses unique challenges: model predictions are usually deterministic functions of input features and highly correlated with outcomes, which can make the causal effects of predictions on outcomes impossible to disentangle from the direct effect of the covariates. We study this problem through the lens of causal identifiability. Despite the hardness of this problem in full generality, we highlight three natural scenarios where the causal effect of predictions can be identified from observational data: randomization in predictions, overparameterization of the predictive model deployed during data collection, and discrete prediction outputs. Empirically we show that given our identifiability conditions hold, standard variants of supervised learning that predict from predictions by treating the prediction as an input feature can find transferable functional relationships that allow for conclusions about newly deployed predictive models. These positive results fundamentally rely on model predictions being recorded during data collection, bringing forward the importance of rethinking standard data collection practices to enable progress towards a better understanding of social outcomes and performative feedback loops.