3 ms·
I don't get how RL can be applied in a domain where there is no simulator. So for customer service, to do RL on real customers... well this sounds like it's go
by sgt101 2y ago
I don't get how RL can be applied in a domain where there is no simulator.
So for customer service, to do RL on real customers... well this sounds like it's going to be staggeringly slow and very expensive in terms of peeved customers.
- bronco21016 2y agoI dunno seems like a decent area to start with. A team that collects detailed data on customer service interactions as well as satisfaction of outcomes should be able to create a decent dataset. Then you grade outputs in a simulator to train the model. No need to train on the fly, at least not at first.
- sgt101 2y agoThere's some challenge there. - service interactions are pretty complicated - the satisfaction is dependent on a lot of factors - customers mostly give 7/10 mostly because giving nps is so unnatural