Evaluation Methods and Measures for Causal Learning Algorithms

The convenient access to copious multifaceted data has encouraged machine learning researchers reconsider correlation-based and embrace the opportunity of causality-based learning, i.e., causal (causal learning). Recent years have, therefore, witnessed great effort in developing algorithms aiming help artificial intelligence (AI) achieve human-level intelligence. Due lack ground-truth data, one biggest challenges current research is algorithm evaluations. This largely impedes cross-pollination AI inference hinders two fields benefit from advances other. To bridge conventional (i.e., based on statistical methods) with Big Data intersection learning), this survey, we review commonly used datasets, evaluation methods, measures for using an pipeline similar learning. We focus fundamental tasks causality-aware tasks. Limitations procedures are also discussed. We, then, examine popular tools/packages conclude primary opportunities benchmarking era Data. survey seeks bring forefront urgency publicly available benchmarks consensus-building standards observational data. In doing so, hope broaden discussions facilitate collaboration advance innovation application