Great work! Just reading your paper, I was thinking how one can formally tune \epsilon and \delta to get certain privacy/accuracy guarantees. Any thoughts?
Deep Learning exploded when open tools made it *easy* I'm extremely excited to share our new @NipsConference paper feat: - Federated Learning - Differential Privacy - Encrypted Data/Models #privacy made... *easy* 
Paper: https://arxiv.org/pdf/1811.04017.pdf … Code: https://github.com/OpenMined/PySyft …pic.twitter.com/iJ29MzHN86
-
-
-
Each published algorithm typically comes with this - the part that is less well understood is what kind of accuracy/speed tradeoff you get with various tunings. Our framework is supposed to make such empirical works *much* easier.
End of conversation
New conversation -
-
-
This tweet made my day! I'm in the process of creating a video on Privacy Matters for Data Science and this paper caught my eye. Thanks!
-
We should chat! :)
End of conversation
New conversation -
-
-
hey,
@ogirardot and@WiemZin presented something very similar at@ScalaIO_FR two weeks ago! See thread https://twitter.com/fanf42/status/1057589606303129600 … and link to slides:https://twitter.com/ogirardot/status/1057617973991608320 …Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
why would **profit-driven** institutions democratize deep learning? Because DL is a black box with too many buttons for the masses to keep pushing until, *fingers-crossed*, something magical happens.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Happy to have contributed a little bit with our code from
@coMindOrg. Keep up the good work!Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Very awesome! I've just finished studying differential privacy
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Sir you are making it look really easy for academic scholars like us....kudos
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
@adversariel@david_madras seems up your alleyThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I know that the design of differential privacy needs to include intended application, e.g., if privacy design includes geospatial resolution. Is there research in privacy of data aggregations as well as individual records? Application, e.g.https://www.bbc.com/news/technology-42853072 …
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.