3 ms·
Sure, lots! We've made lots of progress just in the last decade or two. LSTM, recurrent, and convolutional nets. Slightly older (late 90's), but I think Random
by bkanber 8y ago
Sure, lots! We've made lots of progress just in the last decade or two. LSTM, recurrent, and convolutional nets. Slightly older (late 90's), but I think Random Forests were a pretty significant breakthrough.
Another huge one is Paul Graham's own work using Naive Bayes to filter spam.
- p1esk 8y agoBoth convolutional and recurrent architectures were developed in late 80s. LSTM (improved RNN) was published in 1997.