How to Provide Constraints on Counterfactual Generation
xueyagaga opened this issue · 0 comments
xueyagaga commented
Inspired by DICE (https://github.com/interpretml/DiCE/blob/main/docs/source/notebooks/DiCE_model_agnostic_CFs.ipynb),
how can we specify which features to vary and their permissible ranges for valid counterfactual examples in CARLA?
Looking forward to hearing from you.