I really don't have much faith in things like non-parametric Bayesian methods for, say, computer vision. I don't think that has a future. Natural language processing is the next frontier for deep learning. Everyone agrees that the future is in unsupervised learning. --LeCun展开全文 原微博
某计算神经科学家，模拟过一个10^11大的NN，结果是什么也没发现。🔗 网页链接 Ng和Google去年的大规模NN给我的感觉是：这帮人没idea了，于是就使蛮力，反正做这么大一定能吸引眼球，参数多一定能fit出来点东西。搞研究的别整天想着大规模了，brutal force leads to inelegant solutions。展开全文
Julia这门新语言真有那么神奇？Fast as C, dynamic as Ruby, usable for general programming as Python. easy for statistics as R, natural for string processing as Perl, powerful for linear algebra as Matlab, good at gluing programs together as the shell. 🔗 网页链接展开全文
