He’s making a point about a recent trend towards a new kind of software that’s somewhat broader than what Deep Learning typically refers to (“assembling networks of parameterized functional blocks and by training them from examples using some form of gradient-based optimization”), but it certainly still includes Deep Learning.
He may be right about that, but he’s certainly not saying that Deep Learning itself is actually dead.
DP (or DDP) won't kill DL. They still have particular characteristics to apply to different fields.
I am quite pleased with the turnaround in 2018 of UX. Great boost in automated website building, email marketing and sophisticated technologies for dummies like us. Deep learning happens alongside. We come out of fools paradise somehow in the chase.
I can see differential programming as a natural extension of Deep learning. In my opinion we can't announce the death of deep learning. But in any case, differential programming is a new technique needs to prove its success in the field of AI !
DP (or DDP) won't kill DL. They still have particular characteristics to apply to different fields.
I am quite sure that deep learning is not being dead but is the beginning of new era of Modern Artificial Intelligence. Further work has to be done in order to achieve brain capacity level. Collective intelligence, self-organization, and emergence are some avenues to be explored. Nowadays, it is somewhat hard to know what the future holds.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence