Pedro Domingos

Considering that random forests have many layers and beat deep learning in most applications, maybe we just need to rebrand them as deep forests and they'll be the next big thing.

So everybody is crazy because they prefer Deep Learning rather than random forest even the later being better? Give me a break... Random forest scales much worse than Deep Learning to learning high level features from unstructured data such as image, text, video, etc.

Deep learning is generally better for problems with known a priori structure (vision, language) and worse for those without. Check Kaggle, for example. Everyone should know that.

Let's put this way: for unstructured data (image, video, text, audio, speech), deep learning is unconstestablely better. For tabular data, we have currently a fight in which I concede traditional machine learning may *still* be slightly better, mainly considering its simplicity.

With a sufficiently large datasets and complex features deep learning outperforms on tabular data by miles. Most people don’t have access do that much right tabular data so it might not be as well known.