4 ms·

We taught our platform to learn its own pricing decisions

• rezat 3mo ago
  [flagged]
• Hussain04 3mo 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 3mo ago
    You are right, I will take that think about how to improve it. Thank you very much
    • rezat 3mo ago
      It is now fixed, it rehydrates on each turn
      • Hussain04 3mo ago
        Awesome!