Conference demo summary: differentially private model training with Opacus

[em-blue]

We want to summarise/transcribe people’s talks from the conference to help increase the reach of their talk and make it easier for people to browse the key takeaways from the conference.

Find the schedule of the talk here: https://pricon.openmined.org/agenda And the recording of the talk here: https://www.youtube.com/playlist?list=PLUNOsx6Az_ZGKQd_p4StdZRFQkCBwnaY6

If the speaker is an OpenMined member, feel free to ask them for a copy of their slides - otherwise, you can screenshot key parts from their slides.

[divyadixit]

@em-blue I'd like to give this a try but can't seem to find the recording of this talk in this the playlist.

[cescapa]

@divyadixit you are right! the video can be found here: https://www.youtube.com/watch?v=MWPwofiQMdE&feature=emb_logo if you want to lead, I'd be happy to collaborate with editing

[balapriyac]

That'd be great! As ghost platform allows authors collaborating on an article to joint publish, it'd be great if both of you could collaborate on this demo summary :)

On Sun, 31 Jan 2021, 8:04 am Carlos Escapa, notifications@github.com wrote:

@divyadixit https://github.com/divyadixit you are right! the video can be found here: https://www.youtube.com/watch?v=MWPwofiQMdE&feature=emb_logo if you want to lead, I'd be happy to collaborate with editing

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/OpenMined/writing/issues/139#issuecomment-770314944, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALIW4E3ELK6UFGDPCP563QDS4S6T5ANCNFSM4SGIDSDA .

[divyadixit]

@divyadixit you are right! the video can be found here: https://www.youtube.com/watch?v=MWPwofiQMdE&feature=emb_logo if you want to lead, I'd be happy to collaborate with editing

Thanks a bunch @cescapa! I'll be happy to collaborate with you on editing 💯

[divyadixit]

Assigning the issue to myself based on the conversation with @balapriyac :)

[balapriyac]

As discussed with @divyadixit we found content with considerable overlap on Differentially Private Model Training with Opacus on the OM blog. Hence, marking is done. https://blog.openmined.org/pysyft-opacus-federated-learning-with-differential-privacy/ https://blog.openmined.org/differentially-private-deep-learning-using-opacus-in-20-lines-of-code/