Fictitious Play and Regret Matching (AGT 17)

แชร์
ฝัง
  • เผยแพร่เมื่อ 29 พ.ย. 2024

ความคิดเห็น • 6

  • @jicabe577
    @jicabe577 ปีที่แล้ว

    Thanks a lot! That's the way to explain those topics, with simple numerical examples instead of oscure general formulas. Many thanks, sir!!!

  • @nelsonjoppi
    @nelsonjoppi ปีที่แล้ว +1

    awesome

  • @hyk-f4r
    @hyk-f4r ปีที่แล้ว

    Hi Thanks a lot for these videos, learned a lot. Can you kindly get me reference to any textbooks covering these topics. I am ai researcher and hoping to know these concepts in much more depth.

    • @csprof
      @csprof  ปีที่แล้ว +4

      Unfortunately, there isn't a great textbook for this course, so we had to cobble together several sources. What I wish existed would be an implementation-focused textbook for an upper-level-undergraduate or beginning-graduate CS course. The closest thing to that, and the primary textbook we used for this class is Multiagent Systems by Shoham and Leyton-Brown, which they make available as a free download on their masfoundations website. We also used the undergraduate textbook Essentials of Game Theory by the same authors. Most of the readings I assigned were from those books or from random sites around the web.
      Another useful reference is A Course in Game Theory by Osborne and Rubinstein, which is a proof-heavy textbook for a grad-level econ course in game theory that's freely available on the Osborne-Rubinstein-books website. There's also Twenty Lectures on Algorithmic Game Theory by Roughgarden, a grad-level book from a CS-theory perspective that focuses more on mechanism design than what I'd call AGT. Last and probably least, there's a 2007 collection called Algorithmic Game Theory, which is occasionally useful as a reference, but definitely shouldn't be used as a textbook. It used to be freely available, but I don't think that website exists any more.

    • @hyk-f4r
      @hyk-f4r ปีที่แล้ว +1

      Thanks a lot for the quick reply. You are awesome. Will definitely check those. Your videos are the only comprehensive material I found on the topic and many thanks for that. It meant a lot as I am researching on various imperfect game scenarios (mccfr, deepnash, rebel algorithm etc) and you have given me the helping hand. Subscribed ,please continue your amazing teaching. Looking forward for more. ❤

    • @joaquinvigna6438
      @joaquinvigna6438 ปีที่แล้ว

      Drew Fudenberg and David K. Levine. The Theory of Learning inGames. MIT Press, 1998