Player is loading...
5 minutes Q&A - Privacy by Design Age Verification & Online Child Safety
Dr. Onur Yürüten, Head of Age Assurance Solutions and Senior ML Engineer in Privately
Wednesday, Oct. 11, 2023 · 9:41 a.m. · 07m 40s
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:00
huh i think you know a response vision the just the questions
00:00:04
so from um a liability preview of for this company is easy for
00:00:09
for them the chief to use you always kind of perfect remote will reduce really work you know doing some you should
00:00:16
look to recede from just from the would be to preview
00:00:19
yeah perfect question there and multiple things to to to say there
00:00:24
um the first thing to remember is proportionality
00:00:28
my uh i'm not going to advertise my method
00:00:31
to to for backlog it's because because h. check
00:00:35
eh is not sufficient enough uh to do that
00:00:39
um the way uh the regulators are they are a working more rigorously
00:00:44
to define quantitatively what that a proportion of that means we use cases
00:00:49
uh they are not dividing into basic via a
00:00:53
basic uh standard at a enhanced and and strict levels
00:00:58
of a assurances uh where she uh where we define
00:01:02
the the the price of a miss close classifying someone
00:01:07
uh what is the what what is the uh uh what's the outcome what is the reaction right
00:01:13
uh for instance um yeah that uh of the it
00:01:17
would be the the highest level of security would require too
00:01:21
the to avoid the and the phillies to contact the children but um
00:01:28
playing the playing again a thing again ah at lunchtime
00:01:34
uh maybe you don't need to be that strict uh
00:01:37
it is the heavenly the heavenly uh driven by regulation
00:01:42
in the absence of regulation it's uh it's the profitability
00:01:47
uh that is uh putting question uh and uh and also
00:01:51
we uh the second point i want to bring about is um
00:01:55
uh we usually have this notion of data controller and be in a processor
00:02:00
sh uh in terms of the in the in the in the legal terminology
00:02:05
we already the process and meaning we provide and method
00:02:09
which might be used by our clients uh to get the
00:02:14
result but since it's them who are making a decision they
00:02:17
are the the controller so they are subject to further show
00:02:21
uh for there's a scratchy so so what we
00:02:26
do eh in our conversations is too sensible lies to this subject
00:02:35
okay think to the to the presentation you've
00:02:37
mentioned some tracks um mainly masks whatever part
00:02:43
people's biology is different so some people at each individual
00:02:47
twenty five and some people interactive fire the locate you
00:02:52
and also things like be you read or maybe i know wrinkles or
00:02:57
maybe you start bad to e. to look very good so how do you
00:03:03
how do you figure it out how do you feed those problem because
00:03:07
basically i think your system can be easily gain by just drinking beer
00:03:12
so you can have fourteen year old has you're right people
00:03:16
like yeah yeah so uh so there are multiple layers too
00:03:20
multiple layers to to address this concern uh first of all of after a
00:03:26
certain age after a certain age people uh age very differently uh of shop
00:03:33
so uh if you if you little lives like like mine uh probably you you you are going to be predicted a mature
00:03:39
much older than that someone who's who's living up to you
00:03:42
know half the more more sports uh et cetera right so uh
00:03:47
uh uh we have uh also of in terms of um
00:03:51
a face facial structure yours or your on the spot uh uh
00:03:57
we we uh have done an analysis of uh let's say
00:04:01
skin tones okay we the as fitzpatrick skin tones from from one
00:04:05
to six uh uh and and uh across different age ranges
00:04:10
we we make sure that we make sure that we introduce enough
00:04:15
uh samples in our training shoe to overcome those uh an actual bias
00:04:21
is now coming to the coming to to basic uh forms of attack
00:04:26
well um we do have the uh the the uh the current checks and
00:04:30
balances when we search work points to to avoid a attacks such as you know
00:04:36
crawling point beard on yourself trying to to to put uh other
00:04:40
stuff um but uh but it is uh i feel that keeps evolving
00:04:45
and uh and also we hope we we do have some some cases where
00:04:50
uh where the answer is is really not simple at all uh for instance uh uh what
00:04:55
if someone uh someone has a accessories for religious
00:05:00
reasons they're not willing to remove um so these
00:05:04
are all all some yeah interesting challenges that um that arise and um and in some cases
00:05:10
we we resort to either different climatic signals or
00:05:13
we uh we default to to uh identity check
00:05:17
uh or not and we do that on ah on risk a free space assessment
00:05:23
okay and the second question so how how do you work with their
00:05:27
so say yesterday i was seventeen today's member big tomorrow i'm indian so
00:05:32
my face didn't change a lot about your system pretty probably the same
00:05:37
age but the day before yesterday i wasn't allowed to buy alcohol or cigarettes
00:05:43
but tomorrow i will be hot and you've mentioned retail as part of their things where you would use this so
00:05:50
oh and for people even based down if they see you they would call me
00:05:54
cedar it the day before yesterday they can be fine they can
00:05:57
go to g. or at least it'll make me seekers tomorrow that's okay
00:06:02
so how do you work with beautiful so um
00:06:07
uh so the answer is half technical half the go there a u. k. for instance
00:06:13
they they have defined as something called challenge
00:06:16
twenty five the the clerks they are instructed
00:06:21
the the human say salespeople right they're instructed to definitely ask for i.
00:06:28
d. if they think the person appears to be younger got twenty five
00:06:34
even though even though legally it is possible to sell someone between eighteen
00:06:39
to twenty five so the so the systems are expected to well up old
00:06:45
a a similar snaps standard if not if not better today it today it
00:06:50
would be ridiculous for me to claim that i'll be able to know your age
00:06:55
uh but uh by your will but by precision of a date uh that
00:07:00
maybe there exist such techniques such as uh you know the the scanning the
00:07:04
a bone age uh that which is which is actually something used in medical
00:07:09
uh cases but that but again uh that kind of proper uh precision yes
00:07:14
uh not a proportional to choose certain you stays in cases where it is necessary
00:07:20
to have such a precision done i'm not a i'm not doing its there right now
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
5 minutes Q&A - Privacy-Enhanced Computation in the Age of AI
Dr. Dimitar Jechev, Co-founder and CTO of Inpher
Oct. 11, 2023 · 9:20 a.m.
Privacy by Design Age Verification & Online Child Safety
Dr. Onur Yürüten, Head of Age Assurance Solutions and Senior ML Engineer in Privately
Oct. 11, 2023 · 9:26 a.m.
5 minutes Q&A - Privacy by Design Age Verification & Online Child Safety
Dr. Onur Yürüten, Head of Age Assurance Solutions and Senior ML Engineer in Privately
Oct. 11, 2023 · 9:41 a.m.
(Keynote Talk) Biometrics in the era of AI: From utopia to dystopia?
Dr. Catherine Jasserand, KU Leuven (Belgium), Marie Skłodowska-Curie fellow at Biometric Law Lab
Oct. 11, 2023 · 11:06 a.m.
5 minutes Q&A - Biometrics in the era of AI: From utopia to dystopia?
Dr. Catherine Jasserand, KU Leuven (Belgium), Marie Skłodowska-Curie fellow at Biometric Law Lab
Oct. 11, 2023 · 11:42 a.m.
AI and Privacy
Alexandre Jotterand, CIPP/E, CIPM, attorney-at-law, partner at id est avocats
Oct. 11, 2023 · 11:48 a.m.
5 minutes Q&A - AI and Privacy
Alexandre Jotterand, CIPP/E, CIPM, attorney-at-law, partner at id est avocats
Oct. 11, 2023 · 12:06 p.m.
Preliminary Pperspectives on the Ethical Implications of GenAI
Julien Pache, A Partner at Ethix and Venture Partner at Verve Ventures
Oct. 11, 2023 · 12:12 p.m.
5 minutes Q&A - Preliminary Pperspectives on the Ethical Implications of GenAI
Julien Pache, A Partner at Ethix and Venture Partner at Verve Ventures
Oct. 11, 2023 · 12:30 p.m.
AI & Media: Can You Still Trust Information
Mounir Krichane, Director of the EPFL Media Center
Oct. 11, 2023 · 12:32 p.m.
5 minutes Q&A - AI & Media: Can You Still Trust Information
Mounir Krichane, Director of the EPFL Media Center
Oct. 11, 2023 · 12:54 p.m.
(Keynote Talk) Unlocking the Power of Artificial Intelligence for Precision Medicine with Privacy-Enhancing Technologies
Prof. Jean Louis Raisaro, CHUV-UNIL, assistant professor of Biomedical Informatics and Data Science at the Faculty of Biology and Medicine and the head of the Clinical Data Science Group at the Biomedical Data Science Center
Oct. 11, 2023 · 1:22 p.m.
5 minutes Q&A - Unlocking the Power of Artificial Intelligence for Precision Medicine with Privacy-Enhancing Technologies
Prof. Jean Louis Raisaro, CHUV-UNIL, assistant professor of Biomedical Informatics and Data Science at the Faculty of Biology and Medicine and the head of the Clinical Data Science Group at the Biomedical Data Science Center
Oct. 11, 2023 · 1:50 p.m.
Genomics, AI and Privacy
Julien Duc, Co-Founder and Co-CEO of Nexco Analytics
Oct. 11, 2023 · 2:01 p.m.
5 minutes Q&A - Genomics, AI and Privacy
Julien Duc, Co-Founder and Co-CEO of Nexco Analytics
Oct. 11, 2023 · 2:18 p.m.
How trust & transparency lead the success of an Idiap student's Master's project in fraud detection
Raphaël Lüthi, Machine Learning and AI Lead at Groupe Mutuel
Oct. 11, 2023 · 2:22 p.m.
5 minutes Q&A - How trust & transparency lead the success of an Idiap student's Master's project in fraud detection
Raphaël Lüthi, Machine Learning and AI Lead at Groupe Mutuel
Oct. 11, 2023 · 2:38 p.m.
Recommended talks
Management of health data in an industrial company
V. Srikanth Nallanthighal
Sept. 5, 2019 · 1:57 p.m.
146 views