We taught our platform to learn its own pricing decisions

2 pointsposted 12 hours ago
by rezat

3 Comments

Hussain04

12 hours ago

How does the learner forget old data? If a model's price drops by 50%, or a new model is swapped into an existing rung, the historical weights learned from the old pricing/capability will actively harm routing decisions. Is there a decay factor on the weights or something else?

rezat

12 hours ago

You are right, I will take that think about how to improve it. Thank you very much

rezat

12 hours ago

[flagged]