This talk by Sumedha explains their latest work about explaining black-boxes such as Deep Neural Networks in a smooth way by using counterfactuals and causal analysis.