Player is loading...
5 minutes Q&A - Privacy-Enhanced Computation in the Age of AI
Dr. Dimitar Jechev, Co-founder and CTO of Inpher
Wednesday, Oct. 11, 2023 · 9:20 a.m. · 04m 58s
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:04
i'm just wondering um um you know just as a company now that
00:00:08
the uh you'd have more east to actually get these projects bowling and
00:00:14
implementation how fast now you can actually implement them and have the demand from
00:00:19
from from these different customers did you did you see change
00:00:22
recently or or or not you to regulation or something else so
00:00:26
you're definitely seeing the maturing cover marked things like in general ah
00:00:33
what is still must think is we're getting closer to the regulations
00:00:37
the regulations they're going to create the urgency for the adoption for sure we're
00:00:43
not yet there will still like two years from from from that time that's
00:00:47
what we're we're we're saying and doing these two yes obviously people are first
00:00:52
of all you don't have to explain from scratch to customers what these technologies are
00:00:58
the only become prepare we just got the previous dockets
00:01:01
it but but but was the perfect example are so um
00:01:06
second of all i think the markets tools that's to me
00:01:10
it's a lot of education on actually help extend the value
00:01:15
from that call to convert like our license for such a technology into
00:01:23
oh so so so what's the r. y. mobile essentially probably you probably realise
00:01:27
that the business about that's still what so people know that this is necessary dustin
00:01:33
to get a decay people hall um they could actually uh that
00:01:39
the actions on with these technologies and the more education companies liking for
00:01:46
provide to the mark to the use cases and so on but
00:01:51
but but the foster but could not but that adoption could becomes huh
00:01:58
i will take one last question given up
00:02:02
i i think for much i hope i can ask
00:02:08
too many questions participant interesting presentation it's name price intact
00:02:14
and i have one practical question i'm not an expert in motion correction but when it just
00:02:19
the p. r. it several times and many years ago and and and my experience was that when it
00:02:24
happened to use it for image authentication on button collection
00:02:27
and contact a lot of computational issues competition and learn
00:02:31
any person exceeded maxim was in h. hundred twenty eight hundred twenty eight ah in
00:02:38
all the application especially magic l. and landed talking about f. and that i'd fatal to
00:02:43
de de de de de de date and said that that's a enormous dimensions how
00:02:47
you're handling obvious chaining to say jake an
00:02:50
enormous amount of information especially blind homework construction
00:02:53
that's an excellent question then this exactly uh and brings me to the
00:02:58
point of the hybrid work falls that i was talking about so you in
00:03:03
in such a complex work all your them all ah never going to use
00:03:07
common more fit encryption all all so you but what you would do is
00:03:12
a lot of these little processing you're gonna beat locally
00:03:15
uh because people's in plain text and only then you would
00:03:20
decide what the right data you would feed into the
00:03:24
privacy preserving but it's in contacting them then you are not
00:03:28
only going to use come up a conclusion used for example what
00:03:31
multiparty computation it's kills more and are so common marketing corruption hot
00:03:38
you like they only use it in the model pain for instance
00:03:41
but you could use it in certain aspects of the mobile importance
00:03:45
and part the other also put your question is that mower these they're just being
00:03:51
a lot of initiatives in part are a lot of
00:03:55
work that's the common hardware acceleration out like every every
00:04:00
few months now we're seeing festival new algorithms that enable
00:04:06
us to go in proper selection but we also sink
00:04:09
a lot of our one down by the carpet manufacturers
00:04:13
to ah to talk to enable like intel in the
00:04:17
the uh for instance to to allow for um to
00:04:21
to to basically design that special colour where that could that
00:04:25
potentially ah accelerate common murphy contraption is very interesting art
00:04:30
work that's currently being by most warm optical computing can
00:04:34
pop go fourier transforms that are that was out to be very relevant for some of the building blocks of common
00:04:41
more encryption so we're expecting to see some progress it's
00:04:46
not gonna solve all the problems but in the hybrid
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.
2509 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.
138 views
