thank you thank you and hi everybody I
me and mine who I am and I guess when we
talk about data mining it’s not
necessarily to talk about big data but
we need to stress some information of
what we mean by data mining data mining
the way in which we search is different
from what we are used to do we don’t
search we don’t start from an hypothesis
but we this data mining process search
for a policies that there’s correlation
among subsets of data and those
correlations are somehow summarized in
what is called a pattern and a pattern
is basically a statement which is a sort
of summary of all the instances there’s
led to a sort of conclusion then a model
is created and we are basically fitted
to the model so a lot professor of
giving a talk normally statistically
wears a tie and it’s not dressed this
way so I’m not fitting the model the key
point here is that the pattern is not
truthful across all data so I might not
fit the pattern so I live in a networked
Society so we live basically immersed in
a number of connections all our data are
we produce data data are produced on us
data are referred on X so the part on it
which is created on a number of other
individuals a number of data maybe not
personal identification information are
referred to us no professor there for a
time so we don’t talk about big data we
moved to an interconnection in which
basically all flows of information are
continuously reshaped put
later the pattern is changed we are fit
and refit in the pattern and they are
continuously changing we are even to the
largest and forced to be permanently
connected think about the switch to
clouds services products basically John
allows almost at all to have offline
alternatives and in a sense this is a
way in which the dynamic of our economy
has been moving and transforming
business in permanently connected people
when I talk about permanently connected
I’m not talking about only being
interconnected with Internet I mean that
all our transactions all our
interrelations are mostly computer
mediated so even what we do in our real
life and the physical life it’s computer
mediated think about a credit card
transaction as an example and think
about what we basically do moving around
through all the devices we are using and
all this process is basically enabling
what borrowing from geography we can
call remote sensing so without even
perceiving it data on us our producers
are in further referred we give up data
continuously and there is a permanent
interconnection the results is basically
that we are assuming that we give up
data we don’t even perceive that data
are generated on us without our
contributing to those data think about
behavioral patterns which are supposed
when we are searching the web so we
reached a point in which the economy has
been named by Al Gore the stalker
economy basically it’s the situation in
which businesses
and governments but mostly businesses
know about us much more than we do know
that is from big data analysis pattern
are extracted that actually classify us
according to subsets of data data which
we might not be even aware I’m not aware
they’re moving in a certain way my hands
but the pattern can spot it and act
accordingly I prefer to talk in terms of
the classifying society what is the
classifying society is basically dead
our environment our decision-making
environment is permanently shaped and
reshaped by these flows of information
to think in clear terms all our search
on web are just not objective searches
but are personalized results so we move
we do a search and the results have been
getting or you are getting or different
because they are personalized according
to what the algorithm think beliefs
there is more suitable for me or for you
so our contextual decision-making
environment is continuously changed and
altered in this change so classify
normally meant to a range in category
and Google has clearly changed it and
it’s just an example in classify
arranging people or things never before
was used for people the terms classified
in categories according to shared
qualities or characteristics and this
shared qualities or characteristics
which are gathered which are in third
may not entirely correspond what we
think we are and sometimes they even
take priority I mean in the classifying
society something more tricky if you
want is happening we are used to more
stable and less dynamic classification
white man
white women black men black women in all
the time we have spots that some of them
were too discriminatory so they are not
acceptable in legal or ethical terms it
took a number of years but we are living
now it’s something different in which
the pigeonholing is more transient so
for a number of reasons we are pigeon
holed in a number of cluster according
to a number of data points results we
don’t even know why a part of a woman
would be in a specific cluster getting a
specific context in real and digital
life and this is all changing
continuously adapting we have just heard
about the machine learning and all those
beautiful technologies advances up to
the point that we don’t actually know
and it’s not clearly readable which are
the set of facts the data which gathered
together lead to a pattern a statement
which is user as a knowledge that is to
act in the action it’s actually
surrounding us think about personalized
advertising think about the search and
things similar to that so we are now
living in a world in which the
pigeonholing the classification are more
dynamic continuously change and
transient after five minutes or a few
seconds we can be pigeonholed in
different ways according to different
criteria and those data that pigeon all
may not even be strictly related as we
shall see to us at all so
[Laughter]
I think this basically tell us much more
than a long talk we react to context we
react to stimulus and a number of you
began to laugh or smile just watching
this movies and abstracted from a
commercial so this switch from more
transient dynamic mostly unnoticed not
visible classifications that reshape our
decision-making context are the seminal
switching to more natural life in which
our right to be divergent if you want to
be to react differently from what the
pattern is suggesting you should do
wearing a tie it’s becoming more and
more problematic we’re of course we are
used to systems in which we are somehow
coerced forced to behave in a certain
ways but we are moving away from that
and we are now moving in a completely
different environment in which the shape
of the information the contextual
decision-making is continually
personalised honest and it’s changing so
we switched from just a reshaping of the
amount the quality quantity kind of
information we add or at least we are
thinking we have to what is suggestion
it’s presented as personalization of
services those are the suggestions that
we
should maybe begin to wonder whether we
actually correspond to what this
continuous pattern is suggesting to us
so whether or not we know all of people
or we don’t know any of them the system
and this system confused interaction is
a service allows you to interact with
your friends yet the point is who are my
friends
this is friends for Facebook but these
are subsets of friends for you and you
might not want to mix your family with
college friends with Santana friends and
some of them not even your friends but
still the system is suggesting you that
those people should be your friends
so the impact also in social life is
becoming to be a little bit more complex
and complicated so this has been with us
for a number of years with a lot of
impacts with scoring and other
implications what we are now looking at
is something different once it is not
necessary anymore to identify you as
high who you are Giovanni kimande and
it’s more and more possible to target
individuals or better way to target
devices used or even better groups of
data points of information things begin
to change and all the data which are
referred to us may be somehow misleading
let me give you a may be hilarious
example think about an individual that
for personal reasons for work or any
other reason is somehow forced at the
regular times to go through an area
where the capo prostitutes is going on
and this traffic stop and go
and this process of stopping and go it’s
gathered from the devices mobile phone
tablets computer that brings with him
it’s not interconnected it’s not
connected to internet but those devices
are Jo located temporal is partially
dimensionally located and their movement
is tracked stop and go in it is clearly
a pattern which is compatible with
picking up prostitute what actually then
is the result is that one of those
advices once is connected the
information about those devices those
movements are related to a pattern
result not individual but the set of
devices which have been experiencing
that stop and go at regular chiming are
considered sex addicts or interested in
drugs in specific directions and
dimension well clearly we don’t need to
identify and connect the device with an
individual so for instance the car or
the notebook or the tablet with me or
with a specific gender because in the
internet of things a pink cover can
signal that most probably that
individual is female not male therefore
making it different in the pattern so
not picking up prostitute that would be
the pattern not always true and changing
the results and of course the individual
with the pink cover with not picking up
the prostitute might end up with a wrong
inference of course and bronc
suggestive which of course makes us
laughing and makes us thinking that well
this is not really happening and we do
know that it’s actually happening
already it’s already happening because
for instance cross matching
for devices it’s a huge market in which
basically devices are linked either
because you are connected with your
unique identifier your many unique
identifiers you are connected with your
Apple ID your Google ID or whatever
other ID you are connected or because
your device which is a bundle of other
devices with unique identifiers for
instance the MAC address can be
probabilistically related with each
other and this is something continuously
epping there is plenty of companies
which make a large plus large profits
just making offline and online
identities through probabilistic
matching that trust me works very very
very well now we moved from I I am Joe
Bendigo Monday I but I am supposed to
wear a tie I’m not wearing a tie but the
inference on me is that I should wear a
tie and we reach the point in which
basically it’s not anymore me targeted
in this some devices some objects which
doesn’t really matter whether or not
they are related to me so we are a step
ahead even beyond the idea of data
protection so the frontiers between my
data the data in further me and the data
of the things which some house are
surrounding me are once again sailing
away and and blurring and it’s much much
much easier and effective to target
devices which are matched
among themselves because it’s much more
accurate than targeting me think about
the different objects in the daily life
between 9:30 a.m. and 5:30 and finally
the fitting model they
those devices create is that you are a
professional female age 42 and then
times pass along the day goes on and we
have a different moment in time a
different connection or a different
spatial location of those devices and
clearly some devices come in some other
devices come out and another fitting
model tends to be applied and at that
fitting model in that fitting model you
are a child male age eight just because
maybe your son is using at that regular
time your iPad or your computer and
therefore it’s a specific target
therefore the search ranking will be
tailored for a child male aged eight not
a female 8 42 the results the
suggestions the results in the search
the banners the commercials the filters
and so on so forth these are just
example that if we move along during the
day you might realize at a certain point
the fitting model is not anymore
15:42 but it’s a teenager age 15 right
and if you begin to think about and for
instance this happens only on wetness d
and on every other week on during
weekdays on every other week you can
easily think without being a smart
algorithm Lee you have a shared custody
of the child
so at that point another pattern applies
the child is with mommy or with daddy
and therefore we could somehow
manipulate the environment knowing that
the eight-year-old child is with the
father or with the mother to
contextualise
even more personalizing is
decision-making and environment and
interfering with the decision-making
environment of the adult which is
present at the same time of course times
past alone and
at a certain moment in time the fitting
model is telling to you that you are who
you are and then at the very end of the
day you’d realize that you are your wife
and doesn’t really make a difference
whether or not you are your wives
if the system is contextualizing all
your decision-making environment in that
direction so we reach I guess and I up
the frontiers we want to reach so in the
classifying society in this way in the
way in which our economy has been is
being shaved thanks to technology with
doing Marlo’s things and not saying that
all this is wrong but I’m trying to
stress one point that is we are in a
moment in time and moment in the
development of the economy and
technology in which I our real-life
identities are becoming less important
if you want to a large extent even in
relevant and this irrelevance is
strictly related to the fact that all
the products and services we receive are
contributing to shape our environment in
a specific environment our
decision-making environment so we are at
the point in which we have to decide
whether or not it is legally permissible
or ethically acceptable and to what
extent of withdrawal of information so
the ranking the results which have been
taking away from your search to what
extent this is something we do want and
to what extent withdrawing information
reducing filtering information or
results or possibilities or alternatives
or suggestions Sandman is reaching a
point in which is
too much altering pre defining our
decision-making context it’s how to draw
the border between the marvelous results
of taking away of the noise in the
search to remain in the example and to
not receive any further information on
that well almost twenty five years ago
or even more there was this picture
paper and the idea was on the internet
nobody knows you are dog well with on
sites we might begin to think that
certainly on internet nobody knows you
are a dog but it might be relevant if
you are treated like a dog despite the
fact you might be the owner or you might
not be the owner of a specific kind of
dog and this might be relevant if your
decision making context is shaped by
that well I think these are frontiers
words to be explored
thank you
[Applause]
me and mine who I am and I guess when we
talk about data mining it’s not
necessarily to talk about big data but
we need to stress some information of
what we mean by data mining data mining
the way in which we search is different
from what we are used to do we don’t
search we don’t start from an hypothesis
but we this data mining process search
for a policies that there’s correlation
among subsets of data and those
correlations are somehow summarized in
what is called a pattern and a pattern
is basically a statement which is a sort
of summary of all the instances there’s
led to a sort of conclusion then a model
is created and we are basically fitted
to the model so a lot professor of
giving a talk normally statistically
wears a tie and it’s not dressed this
way so I’m not fitting the model the key
point here is that the pattern is not
truthful across all data so I might not
fit the pattern so I live in a networked
Society so we live basically immersed in
a number of connections all our data are
we produce data data are produced on us
data are referred on X so the part on it
which is created on a number of other
individuals a number of data maybe not
personal identification information are
referred to us no professor there for a
time so we don’t talk about big data we
moved to an interconnection in which
basically all flows of information are
continuously reshaped put
later the pattern is changed we are fit
and refit in the pattern and they are
continuously changing we are even to the
largest and forced to be permanently
connected think about the switch to
clouds services products basically John
allows almost at all to have offline
alternatives and in a sense this is a
way in which the dynamic of our economy
has been moving and transforming
business in permanently connected people
when I talk about permanently connected
I’m not talking about only being
interconnected with Internet I mean that
all our transactions all our
interrelations are mostly computer
mediated so even what we do in our real
life and the physical life it’s computer
mediated think about a credit card
transaction as an example and think
about what we basically do moving around
through all the devices we are using and
all this process is basically enabling
what borrowing from geography we can
call remote sensing so without even
perceiving it data on us our producers
are in further referred we give up data
continuously and there is a permanent
interconnection the results is basically
that we are assuming that we give up
data we don’t even perceive that data
are generated on us without our
contributing to those data think about
behavioral patterns which are supposed
when we are searching the web so we
reached a point in which the economy has
been named by Al Gore the stalker
economy basically it’s the situation in
which businesses
and governments but mostly businesses
know about us much more than we do know
that is from big data analysis pattern
are extracted that actually classify us
according to subsets of data data which
we might not be even aware I’m not aware
they’re moving in a certain way my hands
but the pattern can spot it and act
accordingly I prefer to talk in terms of
the classifying society what is the
classifying society is basically dead
our environment our decision-making
environment is permanently shaped and
reshaped by these flows of information
to think in clear terms all our search
on web are just not objective searches
but are personalized results so we move
we do a search and the results have been
getting or you are getting or different
because they are personalized according
to what the algorithm think beliefs
there is more suitable for me or for you
so our contextual decision-making
environment is continuously changed and
altered in this change so classify
normally meant to a range in category
and Google has clearly changed it and
it’s just an example in classify
arranging people or things never before
was used for people the terms classified
in categories according to shared
qualities or characteristics and this
shared qualities or characteristics
which are gathered which are in third
may not entirely correspond what we
think we are and sometimes they even
take priority I mean in the classifying
society something more tricky if you
want is happening we are used to more
stable and less dynamic classification
white man
white women black men black women in all
the time we have spots that some of them
were too discriminatory so they are not
acceptable in legal or ethical terms it
took a number of years but we are living
now it’s something different in which
the pigeonholing is more transient so
for a number of reasons we are pigeon
holed in a number of cluster according
to a number of data points results we
don’t even know why a part of a woman
would be in a specific cluster getting a
specific context in real and digital
life and this is all changing
continuously adapting we have just heard
about the machine learning and all those
beautiful technologies advances up to
the point that we don’t actually know
and it’s not clearly readable which are
the set of facts the data which gathered
together lead to a pattern a statement
which is user as a knowledge that is to
act in the action it’s actually
surrounding us think about personalized
advertising think about the search and
things similar to that so we are now
living in a world in which the
pigeonholing the classification are more
dynamic continuously change and
transient after five minutes or a few
seconds we can be pigeonholed in
different ways according to different
criteria and those data that pigeon all
may not even be strictly related as we
shall see to us at all so
[Laughter]
I think this basically tell us much more
than a long talk we react to context we
react to stimulus and a number of you
began to laugh or smile just watching
this movies and abstracted from a
commercial so this switch from more
transient dynamic mostly unnoticed not
visible classifications that reshape our
decision-making context are the seminal
switching to more natural life in which
our right to be divergent if you want to
be to react differently from what the
pattern is suggesting you should do
wearing a tie it’s becoming more and
more problematic we’re of course we are
used to systems in which we are somehow
coerced forced to behave in a certain
ways but we are moving away from that
and we are now moving in a completely
different environment in which the shape
of the information the contextual
decision-making is continually
personalised honest and it’s changing so
we switched from just a reshaping of the
amount the quality quantity kind of
information we add or at least we are
thinking we have to what is suggestion
it’s presented as personalization of
services those are the suggestions that
we
should maybe begin to wonder whether we
actually correspond to what this
continuous pattern is suggesting to us
so whether or not we know all of people
or we don’t know any of them the system
and this system confused interaction is
a service allows you to interact with
your friends yet the point is who are my
friends
this is friends for Facebook but these
are subsets of friends for you and you
might not want to mix your family with
college friends with Santana friends and
some of them not even your friends but
still the system is suggesting you that
those people should be your friends
so the impact also in social life is
becoming to be a little bit more complex
and complicated so this has been with us
for a number of years with a lot of
impacts with scoring and other
implications what we are now looking at
is something different once it is not
necessary anymore to identify you as
high who you are Giovanni kimande and
it’s more and more possible to target
individuals or better way to target
devices used or even better groups of
data points of information things begin
to change and all the data which are
referred to us may be somehow misleading
let me give you a may be hilarious
example think about an individual that
for personal reasons for work or any
other reason is somehow forced at the
regular times to go through an area
where the capo prostitutes is going on
and this traffic stop and go
and this process of stopping and go it’s
gathered from the devices mobile phone
tablets computer that brings with him
it’s not interconnected it’s not
connected to internet but those devices
are Jo located temporal is partially
dimensionally located and their movement
is tracked stop and go in it is clearly
a pattern which is compatible with
picking up prostitute what actually then
is the result is that one of those
advices once is connected the
information about those devices those
movements are related to a pattern
result not individual but the set of
devices which have been experiencing
that stop and go at regular chiming are
considered sex addicts or interested in
drugs in specific directions and
dimension well clearly we don’t need to
identify and connect the device with an
individual so for instance the car or
the notebook or the tablet with me or
with a specific gender because in the
internet of things a pink cover can
signal that most probably that
individual is female not male therefore
making it different in the pattern so
not picking up prostitute that would be
the pattern not always true and changing
the results and of course the individual
with the pink cover with not picking up
the prostitute might end up with a wrong
inference of course and bronc
suggestive which of course makes us
laughing and makes us thinking that well
this is not really happening and we do
know that it’s actually happening
already it’s already happening because
for instance cross matching
for devices it’s a huge market in which
basically devices are linked either
because you are connected with your
unique identifier your many unique
identifiers you are connected with your
Apple ID your Google ID or whatever
other ID you are connected or because
your device which is a bundle of other
devices with unique identifiers for
instance the MAC address can be
probabilistically related with each
other and this is something continuously
epping there is plenty of companies
which make a large plus large profits
just making offline and online
identities through probabilistic
matching that trust me works very very
very well now we moved from I I am Joe
Bendigo Monday I but I am supposed to
wear a tie I’m not wearing a tie but the
inference on me is that I should wear a
tie and we reach the point in which
basically it’s not anymore me targeted
in this some devices some objects which
doesn’t really matter whether or not
they are related to me so we are a step
ahead even beyond the idea of data
protection so the frontiers between my
data the data in further me and the data
of the things which some house are
surrounding me are once again sailing
away and and blurring and it’s much much
much easier and effective to target
devices which are matched
among themselves because it’s much more
accurate than targeting me think about
the different objects in the daily life
between 9:30 a.m. and 5:30 and finally
the fitting model they
those devices create is that you are a
professional female age 42 and then
times pass along the day goes on and we
have a different moment in time a
different connection or a different
spatial location of those devices and
clearly some devices come in some other
devices come out and another fitting
model tends to be applied and at that
fitting model in that fitting model you
are a child male age eight just because
maybe your son is using at that regular
time your iPad or your computer and
therefore it’s a specific target
therefore the search ranking will be
tailored for a child male aged eight not
a female 8 42 the results the
suggestions the results in the search
the banners the commercials the filters
and so on so forth these are just
example that if we move along during the
day you might realize at a certain point
the fitting model is not anymore
15:42 but it’s a teenager age 15 right
and if you begin to think about and for
instance this happens only on wetness d
and on every other week on during
weekdays on every other week you can
easily think without being a smart
algorithm Lee you have a shared custody
of the child
so at that point another pattern applies
the child is with mommy or with daddy
and therefore we could somehow
manipulate the environment knowing that
the eight-year-old child is with the
father or with the mother to
contextualise
even more personalizing is
decision-making and environment and
interfering with the decision-making
environment of the adult which is
present at the same time of course times
past alone and
at a certain moment in time the fitting
model is telling to you that you are who
you are and then at the very end of the
day you’d realize that you are your wife
and doesn’t really make a difference
whether or not you are your wives
if the system is contextualizing all
your decision-making environment in that
direction so we reach I guess and I up
the frontiers we want to reach so in the
classifying society in this way in the
way in which our economy has been is
being shaved thanks to technology with
doing Marlo’s things and not saying that
all this is wrong but I’m trying to
stress one point that is we are in a
moment in time and moment in the
development of the economy and
technology in which I our real-life
identities are becoming less important
if you want to a large extent even in
relevant and this irrelevance is
strictly related to the fact that all
the products and services we receive are
contributing to shape our environment in
a specific environment our
decision-making environment so we are at
the point in which we have to decide
whether or not it is legally permissible
or ethically acceptable and to what
extent of withdrawal of information so
the ranking the results which have been
taking away from your search to what
extent this is something we do want and
to what extent withdrawing information
reducing filtering information or
results or possibilities or alternatives
or suggestions Sandman is reaching a
point in which is
too much altering pre defining our
decision-making context it’s how to draw
the border between the marvelous results
of taking away of the noise in the
search to remain in the example and to
not receive any further information on
that well almost twenty five years ago
or even more there was this picture
paper and the idea was on the internet
nobody knows you are dog well with on
sites we might begin to think that
certainly on internet nobody knows you
are a dog but it might be relevant if
you are treated like a dog despite the
fact you might be the owner or you might
not be the owner of a specific kind of
dog and this might be relevant if your
decision making context is shaped by
that well I think these are frontiers
words to be explored
thank you
[Applause]
Please follow and like us: