Smart "Predict, then Optimize"

แชร์
ฝัง
  • เผยแพร่เมื่อ 27 ม.ค. 2025

ความคิดเห็น • 8

  • @yijieyang2035
    @yijieyang2035 ปีที่แล้ว +1

    Nice speech and excellent work!
    I wonder if the framework is suitable for situations when there are many unknown parameters like demand, load, and so on.

  • @matej6418
    @matej6418 2 ปีที่แล้ว +2

    Amazing work. Ran into this same problem many times. Ur generalization is amazing. D. U have any portfolio optimization or some form of knapsack problem example?

    • @adamelmachtoub2140
      @adamelmachtoub2140  2 ปีที่แล้ว +1

      Code with the paper is here which has portfolio opt : github.com/paulgrigas/SmartPredictThenOptimize But this recent work is the place to go and has knapsack I believe: github.com/khalil-research/PyEPO

  • @DarkTobias7
    @DarkTobias7 2 ปีที่แล้ว +1

    Read the paper and this presentation was amazing.
    I was wondering where can I find the codes of the findings (and if you used python)?

    • @tobiaschen6662
      @tobiaschen6662 2 ปีที่แล้ว +1

      It seems like the OR researchers seldom share their code and data. I could understand that this require extra efforts.
      Really appreciate this piece of work, a lot of amazing ideas.

    • @adamelmachtoub2140
      @adamelmachtoub2140  2 ปีที่แล้ว +2

      Code with the paper is here: github.com/paulgrigas/SmartPredictThenOptimize But this recent work is the place to go: github.com/khalil-research/PyEPO

    • @zhangzhiheng8695
      @zhangzhiheng8695 11 หลายเดือนก่อน +1

      @@adamelmachtoub2140 Very nice work! I am a researcher on causality. The prediction part corresponds to the ITE estimation, while the optimization part corresponds to the policy learning. Your work inspires a promising future for causal policy research! Thanks a lot~

    • @zhangzhiheng8695
      @zhangzhiheng8695 11 หลายเดือนก่อน +1

      Sorry did not notice your research until now.