Brendan McMahan - Guarding user Privacy with Federated Learning and Differential Privacy

Поділитися
Вставка
  • Опубліковано 6 лис 2024

КОМЕНТАРІ • 9

  • @MdJuealMia-t3o
    @MdJuealMia-t3o Рік тому

    Nice talk. Its really informative.
    Thanks

  • @simonmasters3295
    @simonmasters3295 Рік тому

    "Think of it as your phone dreaming at night"! Such a pleasant thought, rather than "Imagine your phone analysing your bank account transactions so we can recommend poducts to buy"

  • @taro10h
    @taro10h 4 роки тому +1

    great talk! thank you.

  • @Karim-nq1be
    @Karim-nq1be 4 роки тому +3

    Since we're talking about privacy, are we asking everyone's permission explicitly for this? I really doubt.

    • @jg9193
      @jg9193 3 роки тому +2

      Why would they ask for permission to protect user privacy? The whole point of federated learning is so companies don't have to collect and store your data to make useful products, and the point of differential privacy is to protect your data during that process. Sure, there are still cases where a malicious client or server could potentially access your data or attack the training process, but the alternative is either no user privacy, or no machine learning products

  • @BPP2310
    @BPP2310 4 роки тому +2

    Great talk

  • @hannibalbra2808
    @hannibalbra2808 2 роки тому

    Hello could i have more documents and papers about federal learning please, I am student and i am interested with this

    • @PatatjesDora
      @PatatjesDora 2 роки тому +1

      This is a great paper: Privacy preservation in Distributed Deep Learning: A survey on Distributed
      Deep Learning, privacy preservation techniques used and interesting
      research directions

  • @Netguy09
    @Netguy09 Рік тому

    SOUNDS LIKE A GREAT WAY TO STEAL SOMEONE'S PERSONAL INFORMATION....!