Will there be a "big breakthrough" in spiking neural networks in the next year?

Question

In this interview (https://www.youtube.com/watch?v=4Otcau-C_Yc), Geoff Hinton responded to a question about the next big thing in AI with an answer the *one* potential big thing is a better learning algorithm for spiking neural networks. This question will resolve positively if, a year from now, someone has proposed or I've seen something that qualifies as a big breakthrough in learning algorithms (or results) from spiking neural networks. If I can't come to a binary conclusion about the question, I'll resolve to the probability that best captures my uncertainty about the answer.

Examples of potential big breakthroughs:
* Discovering a learning algorithm for spiking neural networks that appears to have scaling properties and hardware compatibility analogous to current NNs (transformers in particular) with backprop.
* A competitive or state-of-the-art result with spiking NNs on a popular benchmark.

Please leave more ideas for potential qualifying "big breakthroughs" in the comments!

Manifold Markets · Accepted Answer

No — resolved on Jun 22, 2023 by Manifold Markets prediction market.

#	Trader	Total profit
1		Ṁ34
2		Ṁ23
3		Ṁ11
4		Ṁ10
5		Ṁ9

🏅 Top traders

People are also trading

Related questions