Player is loading...
5 minutes Q&A - Genomics, AI and Privacy
Julien Duc, Co-Founder and Co-CEO of Nexco Analytics
Wednesday, Oct. 11, 2023 · 2:18 p.m. · 03m 58s
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:00
questions
00:00:04
let's take but i had to channel next look must have a question of
00:00:11
so i'm interested in understanding how practical y'all handing these given that's
00:00:15
so your companies that pausing and that you can services yes um
00:00:21
guess which hospital sell i don't know private companies and how do you
00:00:25
and sure mike don't test or agencies and so
00:00:29
usually we don't take care of the their storage of
00:00:32
data then we we collaborate with buttons that would take care of storage which are companies with all the um
00:00:39
he bar in all of this legal things already set and then we collaborate with them
00:00:43
for everything that is the processing when it's data to sensitive we have only access to partial information
00:00:50
so we never have access to all the let's see the name and everything so it's true
00:00:54
that in theory which could be um so that which which is fine for the for the role
00:00:59
but we cannot we would do one of the attacked by um that was explained here
00:01:04
um but that that's not of course the goal of of our company want to
00:01:07
the devices and send back results so we really don't keep any of the data
00:01:11
so that's that's the that's the second aspect like we um usually
00:01:15
we have either already processed aggregated information and then we do the analysis
00:01:19
and if we have access to full genome so then we have like you know blind the you know we don't know what would be don't base in
00:01:26
yeah i know i also by means that no no no no no no no no we
00:01:30
don't even because wait for the reason i was then you like really don't want to take
00:01:35
well one anyways whatsoever so we always put in the um in
00:01:39
the the negotiation of the contract that we don't get the data
00:01:43
and we say okay it one month after the the really delivering results then we would just
00:01:47
remove all of the roll daytime everything we keep the the final report if you need it later
00:01:53
but the the raw data is removed if you have any additional analysis
00:01:57
we would have to start over to just think in because it's uh it's and even in the case where you
00:02:02
have to store it that's what i said then we go directly to um specified uh provide actions bits and then
00:02:08
which have all the legal um protect the data and then restored there which it means that information
00:02:14
that you find for example you and i said concert daytime fine mutation online on the back down
00:02:20
you are not allowed to keep this information to get up and for the
00:02:24
future not to do the lecture su and on on exactly that's a hit anything
00:02:29
yeah you know we don't keep it we we give it to with
00:02:31
we really provided of analysis uh solutions so usually it's it's um if
00:02:37
that's a hospital research centre comes to us and and and start to to pitch a project then we discuss how to do the analysis
00:02:43
usually starts from the really very roadway that it's a intrinsic topic typically anachronistic
00:02:49
would start from the beginning to the phone analysis give them the gene set and everything that they are interested in
00:02:54
and uh with the report but we don't do any of the data nor also for
00:02:58
the future it's it's really not in our interest we really try to sell the analysis
00:03:04
now that discussion with uh for example groups like the one from john dewey too
00:03:09
and i like to combine the two right because always really
00:03:12
the yes developing this platform yes to accelerate and if you can
00:03:17
yeah and i went to was directly deployed on the platform that yes yeah no this this
00:03:22
tools we um so we didn't look at it because yet we didn't have to be a
00:03:27
we didn't have to the big analysis very sensitive data that we're in two weeks ago would say
00:03:33
as you it's more um with less sensitive data we work with but yeah definitely we can
00:03:38
work with this uh this tools i mean different people present that um companies also during uh
00:03:43
encryption description in the cloud and then you don't have access to the
00:03:46
data so this this we need we can uh in touch but uh yes
Conference Program
(Keynote Talk) Privacy-Preserving Machine Learning : theoretical and practical considerations
Prof. Slava Voloshynovskiy, University of Geneva, professor with the Department of Computer Science and head of the Stochastic Information Processing group
Oct. 11, 2023 · 7:55 a.m.
2514 views
5 minutes Q&A - Privacy-Preserving Machine Learning : theoretical and practical considerations
Prof. Slava Voloshynovskiy, University of Geneva, professor with the Department of Computer Science and head of the Stochastic Information Processing group
Oct. 11, 2023 · 8:40 a.m.
Enabling Digital Sovereignity With ML
Vladimir Vujovic, Senior Digital Innovation Manager, SICPA
Oct. 11, 2023 · 8:44 a.m.
5 minutes Q&A - Enabling Digital Sovereignity With ML
Vladimir Vujovic, Senior Digital Innovation Manager, SICPA
Oct. 11, 2023 · 8:58 a.m.
Privacy-Enhanced Computation in the Age of AI
Dr. Dimitar Jechev, Co-founder and CTO of Inpher
Oct. 11, 2023 · 9:01 a.m.
139 views
