how fast does need to be an AI agent to be efficient?

Question

I know nowadays agencies are using GPUs in order to accelerate AI, but how fast should be it to be efficient, I mean I know that depends of how large and complex the assignment is but what would be a way to measure its efficiency and what kind of technology(amount of GPUS,RAM,STORAGE) and techniques need to be used in order to get enough efficency?

Any thoughts from experts would be appreciated

Glad that was helpful. If you haven't previously come across them, you will find the concepts of bounded rationality and automata theory quite useful. — DukeZhou, Jun 15 '17 at 19:38
I asked a sort of similar question a while ago, related to DeepBlue and AlphaGo, but hardware and volume limitations don't seem to be very much emphasized currently. — DukeZhou, Jun 15 '17 at 19:47
Wow..Now I got a lot to read and process. Thanks a lot for all of the information , you shared more than expected — Graphs Rodriguez, Jun 15 '17 at 19:52
This got me thinking about the architecture of self-driving cars (how they have to be fully autonomous i.e. can't rely on network connection) and I found a lot of links to the Nvidia Drive PX-series which the manufacturer categorizes as a supercomputer — DukeZhou, Jun 16 '17 at 21:14

DukeZhou · Accepted Answer · 2017-06-15T19:35:09.600

My understanding of efficiency in this context is in regards to optimization of algorithms as opposed to hardware speed, which is more of a "brute force" component. GPUs may be more energy efficient, but this is distinct from linear optimization of algorithms.

In terms of how much processor speed you need to tackle a given problem, that's in the realm of computational complexity theory and analysis of algorithms.

Amount of GPUs, RAM and storage needed to tackle a given problem are purely a function of the complexity of the problem and the efficiency of the algorithms.

how fast does need to be an AI agent to be efficient?

1 Answers1