Kind: captions
Language: en
the following is a conversation with
Tommaso poggio he's the professor at MIT
and as a director of the Center for
brains minds and machines sited over
100,000 times his work has had a
profound impact on our understanding of
the nature of intelligence in both
biological and artificial neural
networks
he has been an advisor to many highly
impactful researchers and entrepreneurs
in AI including demis hassabis of deep
mind I'm nacho of mobile eye and
Christof Koch of the Allen Institute for
brain science this conversation is part
of the MIT course on artificial general
intelligence and the artificial
intelligence podcast if you enjoy it
subscribe on youtube itunes or simply
connect with me on twitter at Lex
Friedman spelled Fri D and now here's my
conversation with Tommaso poggio you've
mentioned that in your childhood you've
developed a fascination with physics
especially the theory of relativity and
that Einstein was also a childhood hero
to you what aspect of Einstein's genius
the nature was genius do you think was
essential for discovering the theory of
relativity you know Einstein was a hero
to me and I'm sure to many people
because he was able to make of course a
major major contribution to physics with
simplifying a bit just a Gedanken
experiment a fourth experiment you know
imagining communication with Lights
between a stationary observer and
somebody on a train and I thought you
know the the fact that just with the
force of his fault of his thinking of
his mind he could guide to some
something so deep in terms of physical
reality how time depend on space and
speed it was something absolutely
fascinating was the power of
intelligence the power of the mind do
you think the ability to imagine to
visualize as he did as a lot of great
forces
sister do you think that's in all of us
human beings or is there something
special to that one particular human
being I think you know all of us can
learn and have in principle similar
breakthroughs
there are lesson to be learned from
Einstein he was one of five PhD students
at ETA and the ID Canarsie technician
actua in Zurich in physics and he was
the worse of the five but the only one
who did not get an academic position
when he graduated well finished his PhD
and he went to work as everybody knows
for the Patent Office and so it's not so
much the work for the Patent Office but
the fact that obviously it was marked
but he was not the top student obviously
was the anti conformist I was not
thinking in the traditional way that
probably stitches and the other students
were doing so there is a lot to be said
about you know trying to be to do the
opposite or something quite different
from what other people are doing that's
actually true for the stock market
never never buy for very bodies by and
also true for science yes so you've also
mentioned staying on a theme of physics
that you were excited and a young age by
the mysteries of the universe that
physics could uncover such as I saw
mentioned the possibility of time travel
so the most out-of-the-box question I
think I'll get to ask today do you think
time travel is possible well it would be
nice if it were possible right now
you know you in science you never say no
but your understanding of the nature of
time yeah it's very likely that it's not
possible to travel in time you may be
able to travel forward in time
if we can for instance freeze ourselves
or you know go on some spacecraft
traveling close to the speed of light
but in terms of activity traveling for
instance back in time I find probably
very unlikely so do you still hold the
underlying dream of the engineering
intelligence that will build systems
that are able to do such huge leaps like
discovering the kind of mechanism that
would be required to travel through time
do you still hold that dream or are
echoes of it from your childhood yeah I
you know I don't think whether there are
certain problems that probably cannot be
solved depending what what you believe
about the physical reality like you know
maybe totally impossible to create
energy from nothing or to travel back in
time but about making machines that can
think as well as we do or better or more
likely especially in the short and
midterm helped us think better which is
in a sense is happening already with the
computers we have and it will happen
more and more but that I certainly
believe and I don't see in principle why
computers at some point could not become
more intelligent than we are although
the word intelligence it's a tricky one
and one who should discuss which I mean
with that in intelligence consciousness
yeah words like love is all these are
very you know you need to be
disentangled so you've mentioned also
that you believe the problem of
intelligence is the greatest problem in
science greater than the origin of life
and the origin the universe you've also
in the talk I've listened to said that
you're open to arguments against against
you so what do you think is the most
captivating aspect of this problem of
understanding the nature of intelligence
why does it captivate you as it does
well originally I think one of the
motivation that I had as I guess a
teenager when I was infatuated with
theory of relativity was really that I I
found that there was the problem of time
and space and general relativity but
there were so many other problems of the
same level of difficulty and importance
that I could even if I were I stein it
was difficult to hope to solve all of
them so what about solving a problem
whose solution allowed me to solve all
the problems and this was what if we
could find the key to an intelligence
you know ten times better or faster than
Einstein so that's sort of seeing
artificial intelligence as a tool to
expand our capabilities but is there
just an inherent curiosity in you and
just understanding what is in our in
here that makes it all all work yes
absolutely all right so I was starting I
started saying this was the motivation
when I was a teenager but you know soon
after I think the problem of human
intelligence became a real focus of you
know of my sent my science and my
research because I think he's for me the
most interesting problem is really
asking oh we we are right is asking not
only a question about science but even
about the very tool we are using to do
science which is our brain how does our
brain work from where does it come from
after its limitation can we make it
better
and that in many ways is the ultimate
question that underlies this whole
effort of science so you've made
significant contributions in both the
science of intelligence and the
engineering event
in a hypothetical way let me ask how far
do you think we can get in creating
intelligent systems without
understanding the biological the
understanding how the human brain
creates intelligence
put another way do you think we can
build a strong-ass system without really
getting at the core the functionally
understanding the functional nature of
the brain well this is a real difficult
question you know we did solve problems
like flying without really using too
much our knowledge about how birds fly
it was important I guess to know that
you could have things heavier than than
air being able to fly like like birds
but beyond that probably we did not
learn very much you know some you know
the brothers right did learn a lot of
observation about birds and designing
their their aircraft but you know you
can argue we did not use much of biology
in that particular case now in the case
of intelligence I think that it's it's a
bit of a bet right now
if you are if you ask okay we we all
agree we'll get at some point maybe soon
maybe later to a machine that is
indistinguishable from my secretary say
in terms of what I can ask the machine
to do I think we get there and now the
question is and you can ask people do
you think we'll get there without any
knowledge about you know the human brain
or that is the best way to get there is
to understand better the human brain
yeah okay this is I think an educated
bet that different people with different
background will decide in different ways
the recent history of the progress in AI
in the last out say five years or ten
years is has been that the main
breakthroughs
the main recent breakthroughs I really
start from neuroscience
mention reinforcement learning as one is
one of the algorithms at the core of
alphago which is the system that beat
the kind of an official world champion
of go lee sedol and two three years ago
in seoul that's one and that started
related with the work of Pavlov and I'll
or hundred Marvin Minsky in the sixties
many other neuroscientists later on and
deep learning started which is the core
again of alphago and systems like
autonomous driving systems for cars like
the systems that mobile I which is a
company started by one of my exposed or
Colonel Joshua did so that is a core of
those things and deep learning really
the initial ideas in terms of the
architecture of this layered ARCIC on
networks started with work of Torsten
Wiesel and David Hubel at Harvard up the
river in the 60s so recent history
suggests the neuroscience played a big
role in these breakthroughs my personal
bet is that there is a good chance they
continue to play a big role maybe not in
all the future breakthroughs but in some
of them at least in inspiration so at
least in a new spirit absolutely yes so
you see you studied both artificial and
biological neural networks you said
these mechanisms that underlie deep
learning deeper and reinforcement
learning but there is nevertheless
significant differences between
biological and artificial neural
networks as they stand now so between
the two what he finds the most
interesting mysterious maybe even
beautiful difference as it currently
stands in our understanding I must
confess that until recently I found that
the artificial networks
too simplistic relative to real neural
networks but you know recently I've been
started to think that yes there are a
very big simplification of what you find
in the brain but on the other hand there
are much closer in terms of the
architecture to the brain than other
models that we had that computer science
used as model of thinking which were
mathematical logics you know Lisp Prolog
and those kind of things yeah so in
comparison to those they're much closer
to the brain you have networks of
neurons which is what the brain is about
and the artificial neurons in the models
are as I said caricature of the
biological neurons but they're still
neurons single units communicating with
other units something that is absent in
you know the traditional computer type
models of mathematics reasoning and so
on so what aspect is would you like to
see in artificial neural networks added
over time as we try to figure out ways
to improve them so one of the main
differences and you know problems in
terms of deep learning today and it's
not only deep learning and the brain is
the need for deep learning techniques to
have a lot of labeled examples you know
for Easter for imagenet you have like a
training site which is 1 million images
each one labeled by some human in terms
of which object is there and it's it's
clear that in biology a baby may be able
to see million of images in the first
years of life but will not have million
of labels given to him or her by parents
or take
take care takers so how do you solve
that you know I think there is this
interesting challenge that today deep
learning and related techniques are all
about big data big data meaning a lot of
examples labeled by humans
whereas in nature you have so that this
big data is n going to infinity that's
the best you know and meaning labeled
data but I think the biological world is
more n going to one Hey a child can
learn the beautiful wrote a very small
number of you know labeled examples like
you tell a child this is a car you don't
need to say like imagenet you know this
is a car this is a car this is not a car
this is not a cat 1 million times so and
of course with alphago and or at least
alpha 0 variants there's because of the
because the world of go is so simplistic
that you can actually learn by yourself
through self play you could play against
each other and the real world i meet the
visual system that you've studied
extensively is a lot more complicated
than the game of go so under comment
about children which are fascinatingly
good at learning new stuff how much of
it do you think is hardware how much of
it is software you know that's a good
deep question is in a sense is the old
question of nurture and nature how much
isn't in the gene and how much is in the
experience of an individual obviously
it's both that play a role and i believe
that the way
evolution gives put prior information so
to speak hard while it's not really hard
while but that's essentially an
hypothesis I think what's going on is
that evolution as you know almost
necessarily if you believe in Darwin is
very opportunistic and and think about
our DNA and the DNA of Drosophila our
DNA does not have many more genes than
resolve around the fly the fly the fruit
fly now we know that the fruit fly does
not learn very much during its
individual existence it looks like one
of this machinery that it's really
mostly not hundred percent but you know
95 percent hard coded by the genes
but since we don't have many more genes
than Drosophila as evolution could
encoding as a kind of general learning
machinery and then had to give very weak
priors like for instance let me take
give a specific example which is recent
to work by a member of our Center for
brains minds and machines we know
because of work of other people in our
group and other groups that there are
cells in a part of our brain neurons
that are tuned to phases they seems to
be involved in face recognition now this
face area exists seems to be present in
young children and adults and one
question is is there from the beginning
is hardwired by evolution or you know
somehow is learned very quickly so
what's your by the way a lot of the
questions I'm asking with the answer is
we don't really know but as a person who
has contributed some profound ideas in
these fields you're a good person to
guess at some of these so of course
there's a caveat before a lot of the
stuff we talk about but what is your
hunch is the face the part of the brain
that that seems to be concentrated on
face recognition are you born with that
or you just is designed to learn that
quickly like the face of the mother and
I my hand shimmer by bias was the second
one learned very quickly and it turns
out that Marge Livingstone at Harvard
has done some amazing experiments in
which she raised baby monkeys depriving
them of faces during the first weeks of
life so they see technicians but the
technician have a mask yes and and so
when they looked
at the area in the brain of this monkeys
that were usually find faces they found
no face preference so my guess is that
what evolution does in this case is
there is a plastic Canaria which is
plastic which is kind of predetermined
to be imprinted very easily but the
command from the gene is not detailed
circuitry for a face template could be
but this will require probably a lot of
bits you had to specify a lot of
connection of a lot of neurons instead
that the command that commands from the
gene is something like imprint memorized
what you see most often in the first two
weeks of life especially in connection
with food and maybe nipples I don't
write well source of food and so in then
that area is very plastic at first and
in the otherwise I'd be interesting if a
variant of that experiment would show a
different kind of pattern associated
with food than a face pattern well
whether that quite stick there are
indications that during that experiment
what the monkey saw quite often where
the blue gloves of the technicians that
were giving to the baby monkeys the milk
and some of the cells see instead of
being face sensitive in that area or a
hand sensitive that's fascinating
can you talk about what are the
different parts of the brain and in your
view sort of loosely and how do they
contribute to intelligence do you see
the brain as a bunch of different
modules and they together come in the
human brain to create intelligence or is
it all one mush of the same kind of
fundamental architecture yeah that's you
know that's an important question and
there was a phase in neuroscience by
in the 1950 or so in which it was
believed for a while that the brain was
equipotential this was the term you
could cut out a piece and nothing
special happened apart a little bit less
performance there was a a surgeon
Lashley did a lot of experiments of this
type with mice and rats and concluded
that every part of the brain was
essentially equivalent to any other one
it turns out that that's that's really
not true it's there are very specific
modules in the brain as you said and you
know people may lose the ability to
speak if you have a stroke in a certain
region or may lose control of their legs
in another region or so they're very
specific the brain is also quite
flexible and redundant so often it can
correct things and you know the kind of
takeover functions from one part of the
brain to the other but but but really
there are specific modules of the answer
that we know from this old work which
was basically on based on lesions either
on animals or very often there were a
mine of well it there was a mine a very
interesting data coming from from the
war from different types of injuries
injuries that soldiers had in the brain
and more recently functional MRI which
allow you to to check which part of the
brain are active when you are doing
different tasks as you know can replace
some of this you can see that certain
parts of the brain are involved or
active in this language yeah yeah that's
right
but sort of taking a step back to that
part of the brain that discovers that
specializes in the face and how that
might be learned what's your intuition
behind you you know is it possible that
the sort of from a physicists
perspective when you get lower and lower
that it's all the same stuff and it just
when you're born it's plastic and it
quickly figures out this part is going
to be about vision this is gonna be
about language this is about common
sense reasoning do you have an intuition
that that kind of learning is going on
really quickly or is it really kind of
solidified in hardware that's a great
question so there are parts of the brain
like the cerebellum or they put campus
that are quite different from each other
they clearly have different Anatomy
different connectivity that then there
is the cortex which is the most
developed part of the brain in humans
and in the cortex you have different
regions of the cortex that are
responsible for vision for audition for
motor control for language now one of
the big puzzles of of this is that in
the cortex is the cortex is the cortex
it looks like it is the same in terms of
hardware in terms of type of neurons and
connectivity across these different
modalities so for the cortex letting
aside these other parts of the brain
like spinal cord upon campus or bedroom
and so on for the cortex I think your
question about hardware and software and
learning and so on it's it I think is
rather open and you know it I find very
interesting for easy to think about an
architecture computer architecture that
is good for vision and the symptom is
good for language seems to be you know
so different problem areas that you have
to solve but the underlying mechanism
might be the same that's really
instructive for it maybe artificial
neural networks so you've done a lot of
great work in vision and human vision
computer vision and you mentioned the
problem of human vision is really as
difficult as the problem of general
intelligence and maybe that connects to
the cortex discussion can you describe
the human visual cortex
and how the humans begin to understand
the world through the raw sensory
information the woods for folks enough
familiar especially in on the computer
vision side we don't often actually take
a step back except saying what the
sentence or two that one is inspired by
the other well what is it that we know
about the human visual cortex that's
interest so we know quite a bit at the
same time we don't know a lot but the
the bit we know you know in a sense we
know a lot of the details and Men we
don't know and we know a lot of the top
level the answer the top level question
but we don't know some basic ones even
in terms of general neuroscience
forgetting vision you know why do we
sleep it's such a basic question and we
really don't have an answer to that do
you think so taking a step back on that
so sleep for examples fascinating do you
think that's a neuroscience question or
if we talk about abstractions what do
you think is an interesting way to study
intelligence or are most effective on
the levels of abstractions the chemicals
the biological is electro physical
mathematical as you've done a lot of
excellent work on that side which
psychology is sort of like at which
level of abstraction do you think well
in terms of levels of abstraction I
think we need all of them all hits when
you know it's like if you ask me what
does it mean to understand the computer
right that's much simpler but in a
computer I could say well I understand
how to use PowerPoint that's my level of
understanding a computer it's it has
reasonable you know give me some power
to produce lights and beautiful slides
and now the class on body exercise well
I I know how the transistor work that
are inside the computer I can write the
equation for you know transistor and
diodes and circuits logical circuits
and I can ask this guy do you know how
to operate PowerPoint no idea so do you
think if we discovered computers walking
amongst us full of these transistors
that are also operating under windows
and have PowerPoint do you think it's
digging in a little bit more how useful
is it to understand the transistor in
order to be able to understand
PowerPoint and these higher-level very
good intelligence I see so I think in
the case of computers because they were
made by engineers by us this different
level of understanding are rather
separate on purpose you know you there
are separate modules so that the
engineer that designed the circuit for
the chips does not need to know what
power is inside PowerPoint and somebody
you can write the software translating
from one to the end to the other and so
in that case I don't think understanding
the transistor help you understand
PowerPoint or very little if you want to
understand the computer this question
you know I would say you have to
understanding a different levels if you
really want to build one right but but
for the brain I think these levels of
understanding so the algorithms which
kind of computation you know the
equivalent of PowerPoint and the
circuits you know the transistors I
think they are more much more
intertwined with each other there is not
you know in Italy level of the software
separate from the hardware and so that's
why I think in the case of the brain a
problem is more difficult or more than
four computers requires the interaction
the collaboration between different
types of expertise that's a big the
brain is a big mess
you can't just on disentangle a level I
think you can but is is much more
difficult and it's not you know it's not
completely obvious and
and I said I think he's one of the
person everything is the greatest
problem in science so yeah you know I
think he's it's fair that it's difficult
one that said you do talk about
compositionality and why I might be
useful and when you discuss what why
these neural networks in artificial or
biological sense learn anything you talk
about compositionality see there's a
sense that nature can be disentangled
our purpura
well all aspects of our cognition could
be disentangled a little to some degree
so why do you think what first of all
how do you see compositionality and why
do you think it exists at all in nature
it spoke about I use the the term
compositionality
when we looked at deep neural networks
multi-layers and trying to understand
when and why they are more powerful than
more classical one layer network like
linear classifier kernel machines
so-called and what we found is that in
terms of approximating or learning or
representing a function a mapping from
an input to an output like from an image
to the label in the image if this
function as a particular structure then
deep networks are much more powerful
than shallow networks to approximate the
underlying function and the particular
structure is a structure of
compositionality if the function is made
up of functions of function so that you
need to look on when you are
interpreting an image classifying an
image you don't need to look at all
pixels at once but you can compute
something from small groups of pixels
and then you can compute something on
the output of this local computation and
so on that is similar to what you do
when you read the sentence you don't
need to read the first and the last
letter but you can read syllables
combine them in words combine the words
in sentences so this is this kind of
structure so that's as part of the
discussion of why deep neural networks
may be more effective than the shallow
methods and is your sense for most
things we can use neural networks for
those problems are going to be
compositional in nature like like
language like vision how far can we get
in this kind of right
so here is almost philosophy well you
know there yeah let's go there so a
friend of mine max tegmark who is a
physicist at MIT I've talked to him on
this thing yeah and he disagrees with
you right yeah but we you know we agree
most but the conclusion is a bit
differently
he is conclusion is that for images for
instance the compositional structure of
this function that we have to learn or
to solve these problems comes from
physics comes from the fact that you
have local interactions in physics
between atoms and other atoms between
particle of matter and other particles
between planets and other planets
between stars that it's all local and
that's true but you could push this
argument a bit further not this argument
actually you could argue that you know
maybe that's part of the true but maybe
what happens is kind of the opposite is
that our brain is wired up as a deep
network so it can learn understand solve
problems that I have this compositional
structure and I cannot do they cannot
solve problems that don't have this
compositional stretch so the problem is
we are accustomed to we think about we
test our algorithms on our this
compositional structure because our
brain is made up in that's in a sense an
evolutionary perspective as we've so the
ones that didn't have the they weren't
dealing with a compositional nature of
reality died off yes it also could be
may be the reason
why we have this local connectivity in
the brain like simple cells in cortex
looking only the small part of the B
image each one of them and another says
looking at it small number of these
simple cells and so on the reason for
this may be purely that was difficult to
grow longer range connectivity so
suppose it's you know for biology it's
possible to grow short range
connectivity but not longer and also
because there is a limited number of
long range the Duke and so you have at
this
this limitation from the biology and
this means you build a deep
convolutional neck this would be
something like deep convolutional
network and this is great for solving
certain class of problem these are the
ones we are we find easy and important
for our life and yes they were enough
for us to survive and and you can start
a successful business on solving those
problems right mobile a driving is a
compositional problem right so on the
unlearning task i mean we don't know
much about how the brain learns in terms
of optimization but so the thing that's
stochastic gradient descent is what
artificial neural networks used for the
most part to adjust the parameters in
such a way that it's able to deal based
on the label data it's able to solve the
problem yeah so what's your intuition
about why it works at all a heart of a
problem it is to optimize in your own
network artificial neural network is
there other alternatives
you're just in general your intuition is
behind this very simplistic algorithm
that seems to do pretty good surprising
yes yes so I find near of science the
the architecture of cortex
it's a really similar to the
architecture of deep networks so that
there is a nice correspondence there
between the biology and this kind of
local connectivity hierarchical
architecture the stochastic gradient
descent as you said is is a very simple
technique
it seems pretty unlikely that biology
could do that
from from what we know right now about
you know cortex and neurons and synapses
so it's a big question open whether
there are other optimization learning
algorithms that can replace stochastic
gradient descent and my my guess is yes
but nobody has found yet a real answer I
mean people are trying still trying and
there are some interesting ideas the
fact that stochastic gradient descent is
so successful this has become clear is
not so mysterious and the reason is that
it's an interesting fact you know it's a
change in a sense in how people think
about statistics and and this is the
following is that typically when you had
data and you had say a model with
parameters you are trying to fit the
model to the data you know to fit the
parameter typically the kind of kind of
crowd wisdom type idea was you should
have at least you know twice the number
of data than the number of parameters
you maybe 10 times is better now the way
you train neural net or this disease
that I have they have 10 or 100 times
more parameters than did exactly the
opposite and which you know it is it has
been one of the puzzles about neural
networks how can you get something that
really works when you have so much
freedom in its in from that Laura Derek
in general right somehow right exactly
do you think this the stochastic nature
is essential to randomness so I think we
have some initial understanding why this
happens but one nice side effect of
having this over parameterization more
parameters than data is that when you
look for the minima of a loss function
like stochastic gradient descent is
doing in find I I made some calculations
based on
some old basic theorem of algebra called
bazoo theorem and that gives you an
estimate of the number of solution of a
system of polynomial equation anyway the
bottom line is that there are probably
more minima for a typical deep networks
than atoms in the universe just to say
there are lost because of the over
parametrization
a more global minimum zero meaning good
meaning so it's not just local minima
yeah a lot of them so you have a lot of
solutions so it's not so surprising that
you can find them relatively easily and
this is why this is because of the
overall parameterization the
organization sprinkles an entire space
for solutions pretty good and so not so
surprising right is like you know if you
have a system of linear equation and you
have more unknowns than equations then
you have we know you have an infinite
number of solutions and the question is
to pick one that's another story but you
have an infinite number of solutions so
there are a lot of value of your
unknowns that satisfy the equations but
it's possible that there's a lot of
those solutions that aren't very good
what's surprising so that's a good
question why can you pick one the
generalizes one yeah that's a separate
question with separate answers one one
theorem that people like to talk about
that kind of inspires imagination of the
power in your networks is the
universality a universal approximation
theorem you can approximate any
computable function with just a finite
number of neurons and a single hidden
layer see you find this theorem one
surprising you find it useful
interesting inspiring now this one you
know I never found it very surprising
it's was known since the 80s
since I entered the field because it's
basically the same as biased as the
which says that I can approximate any
continuous function with a polynomial of
sufficiently with a sufficient number of
terms monomials so basically the same
and the proves very similar so your
intuition was there's never any doubt in
your networks in theory could the right
be very strong approximate nicely the
the question the interesting question is
that if this theorem
it says you can approximate fine but
when you ask how many neurons for
instance or in the case of polynomial
how many monomials I need to get a good
approximation then it turns out that
that depends on the dimensionality of
your function how many variables you
have but it depends on the
dimensionality of your function in a bad
way it's for instance suppose you want
an error which is no worse than 10% in
your approximation you come up with a
net of the approximate your function
within 10% then turns out that the
number of units you need are in the
order of 10 to the dimensionality D how
many variables so if you have you know
two variables is these 2 would you have
hundred units and okay but if you have
say 200 by 200 pixel images now this is
you know 240 thousand whatever and we
can go to the sizing universe pretty
quickly there are exactly 10 to the
40,000 and so this is called the curse
of dimensionality not you know quite
appropriate and the hope is with the
extra layers you can remove the curse
what we proved is that if you have deep
layers or a rocky core architecture that
with the local connectivity of the type
of convolutional deep learning and if
you are dealing with a function that has
this kind of hierarchical architecture
then you avoid completely the curves
you've spoken a lot about supervised
deep learning yeah what are your
thoughts hopes views on the challenges
of unsupervised learning with the
with Ganz with the generator valor
surround networks do you see those is
distinct that the power of Ganz does is
distinct from supervised methods in your
networks are they really all in the same
representation ballpark gains is one way
to get estimation of probability
densities which is somewhat new way but
people have not done before I I don't
know whether this will really play an
important role in you know in
intelligence or it's it's interesting
I'm I'm less enthusiastic about it too
many people in the field I have the
feeling that many people in the field
are really impressed by the ability to
of producing realistic looking images in
this generative way which describes the
popularity of the methods but you're
saying that while that's exciting and
cool to look at it may not be the tool
that's useful for yeah for so you
describe it kind of beautifully current
supervised methods go and to infinity in
terms of number of labelled points and
we really have to figure out how to go
to and to one yeah and you're thinking
ganz might help but they might not be
the right I don't think you for that
problem which I really think is
important I think they may help they
certainly have applications for instance
in computer graphics and you know we I
did work long ago
which was a little bit similar in terms
of saying okay 11 network and I present
images and I can so input its images and
output is for instead the pose of the
image you know a face how much is miling
is rotated 45 degrees or not what about
having a network that I trained with the
same dataset but now I invert input and
output now the input is the pose or the
expression number certain numbers and
the output is the image and I train it
and we did pretty good interesting
results in terms of producing very
realistic looking images was you know
less sophisticated mechanism but the
output was pretty less than gains but
the output was pretty much of the same
quality so I think for computer graphics
type application yeah definitely gains
can be quite useful and not only for
that--for but for you know helping for
instance on this problem of unsupervised
example of reducing the number of
labeled examples I think people it's
like they think they can get out more
than they put in you know it there's no
free lunches Yeah right
that's what do you think what's your
intuition how can we slow the growth of
n to infinity in supervised and to
infinity in supervised learning so for
example mobile I has very successfully I
mean essentially annotated large amounts
of data to be able to drive a car now
one thought is so we're trying to teach
machines of AI and we're trying to so
how can we become better teachers maybe
that's one one way now I got your you
know what I like that because one
again one caricature of the history of
computer sites you could say is with the
gains with programmers expensive yeah
continuously labelers cheap yeah and the
future would be schools like we have for
kids
yeah currently the labeling methods were
not selective about which examples we we
teach networks with so I think the focus
of making one-shot networks that learn
much faster is often on the architecture
side but how can we pick better examples
with wish to learn do you have
intuitions about that well that's part
of the quarter program but the other one
is you know if we look at biology
reasonable assumption I think is in the
same spirit II that I said evolution is
opportunistic and has weak priors you
know the way I think the intelligence of
child the baby may develop is by
bootstrapping weak priors from evolution
for instance in you can assume that you
are having most organisms including
human babies built in some basic
machinery to detect motion and relative
motion and in fact there is you know we
know all insects from fruit flies other
animals they have this
even in the readiness of in the very
peripheral part it's very conserved
across species something that evolution
discovered early it may be the reason
why babies tend to look in the first few
days to moving objects and not to not
moving out now moving objects means okay
they are attracted by motion but motion
also means that motion gives automatic
segmentation from the background so
because of motion boundaries you know
either the object is moving or the eye
of the baby is tracking the moving
object and the background is moving
right yeah so just purely on the visual
characteristics of the scene as seems to
be the most useful right so it's like
looking at an object without background
it's ideal for learning the object
otherwise it's really difficult because
you have so much stuff so suppose you do
this at the beginning first weeks then
after that you can recognize the object
now they're imprinted a number of even
in the background even without motion so
that's at the by the way I just want to
ask an object recognition problem so
there is this being responsive to
movement and edge detection essentially
what's the gap between being effectively
effective at visually recognizing stuff
detecting word that is and understanding
the scene there is this a huge gap in
many layers or is it as a close no I
think that's a huge gap
I think present algorithm with all the
success that we have and the fact that
are a lot of very useful it's I think we
are we are in a golden age for
applications of low level vision and low
level speech recognition and so on you
know Alexa and so
there are many more things of similar
level to be done including medical
diagnosis and so on but we are far from
what we call understanding of a scene of
language of actions of people that is
despite the claims that's I think very
far or a little bit off so in popular
culture and among many researchers some
of which I've spoken with the sue
Russell and you know a mask in and out
of the AAI field there's a concern about
the existential threat of AI yeah and
how do you think about this concern in
and is it valuable to think about
large-scale long-term unintended
consequences of intelligent systems we
try to build I always think is better to
worry first you know early rather than
late so some worry is good yeah I'm not
against worry at all
personally I think that you know it will
take a long time before there is real
reason to be worried but as I said I
think it is good to put in place and
think about possible safety against what
I find a bit misleading are things like
that I've been said by people I know
like Elon Musk and what is boström
important notice first name a neck panic
poster right you know and a couple of
other people that for instance a eyes
more dangerous the nuclear weapons right
yeah I think that's really project that
can be it's misleading because in terms
of priority which should still be more
worried about nuclear weapons and you
know what people are doing about it and
some then a
and he's spoken about them as obvious
and yourself saying that you think
you'll be about a hundred years out
before we have a general intelligence
system that's on par with the human
being you have any updates for those
predictions what I think he said he's at
28 he said it went all right this was a
couple of years ago I have not asked him
again so I should have your own
prediction what's your prediction about
when you'll be truly surprised
and what's the confidence interval or
not you know it's so difficult to
predict the future and even the presence
of it's nothing it's pretty hard to
predict a bit I'll be but as I said this
is completely it would be more like rod
Brooks I think he's about 200 years when
we have this kind of a GI system
artificial general intelligence system
you're sitting in a room with her him it
do you think it will be the underlying
design of such a system is something
we'll be able to understand it will be
simple do you think you'll be
explainable understandable by us your
intuition again we're in the realm of
philosophy a little bit but probably no
but it again it depends
would you really mean for understanding
so I think
you know we don't understand what how
deep networks work I think we're
beginning to have a theory now but in
the case of deep networks or even in the
case of the simple simpler kernel
machines or linear classifier we really
don't understand the individual units
also we but we understand you know what
the computation and the limitations and
the properties of it are it's similar to
many things you know we what does it
mean to understand how a fusion bomb
works how many of us you know many of us
understand the basic principle and some
of us may understand deeper details in
that sense understanding is as a
community as a civilization can we build
another copy of it okay and in that
sense you think there'll be there will
need to be some evolutionary component
where it runs away from our
understanding or do you think it could
be engineered from the ground up the
same way you go from the transistor to
our point
all right so many years ago this was
actually 40 41 years ago I wrote a paper
with David Marr who was one of the
founding father of computer vision of
computational dish I wrote a paper about
levels of understanding which is related
to the question I discussed earlier
about understanding power point
understanding transistors and so on and
you know in that kind of framework we
had the level of the hardware and the
top level of the algorithms we did not
have learning recently I updated adding
levels and one level I added to those
free was learning so and you can imagine
you could have a good understanding
of how you construct learning machine
like we do but being unable to describe
in detail what the learning machines
will discover right now that would be
still a powerful understanding if I can
build the learning machine even if I
don't understand in detail every time
made it learn something just like our
children if they're if they start
listening to a certain type of music I
don't know Miley Cyrus or something you
don't understand why they came after
that particular preference but you
understand the learning process that I'm
very interesting yeah yeah so unlearning
for systems to be part of our world it
has a certain one of the challenging
things that you've spoken about is
learning ethics learning yeah morals and
what how hard do you think is the
problem of first of all humans
understanding our ethics what is the
origin and the neural a low level of
ethics what is it at a higher level is
it something that's learner before
machines in your intuition I think yeah
ethics is learnable very likely I I
think I is one of these problems were
think understanding the neuroscience of
ethics you know people discuss there is
an ethics of neuroscience yes you know
how a neuroscientist should or should
not behave can you think of a
neurosurgeon and the ethics are you Rory
has to behavior he she has to be but I'm
more interested on the neuroscience of
you blow my mind right now the
neuroscience of ethics is very matter
yeah and you know I think that would be
important to understand also for being
able to to design machines that have
that are ethical machines in
our sense of ethics and you think there
is something in your science there's
patterns tools in your science that can
help us shed some light on ethics or
yeah mostly on the psychology sociology
much higher level no there is a culture
but there is also in the meantime there
are there is evidence fMRI of specific
areas of the brain that are involved in
certain ethical judgment and not only
this you can stimulate those area with
magnetic fields and change the ethical
decisions yeah Wow so that's work by a
colleague of mine Rebecca Saxe and there
is a other researchers doing similar
work and I think you know this is the
beginning but ideally at some point
we'll have an understanding of how this
works
and white of all right the big y
question yeah it must have some some
purpose yeah obviously test you know
some social purpose is is probably if
neuroscience holds the key to at least
eliminate some aspect of ethics that
means it could be a learn about problem
yeah exactly
and as we're getting into harder and
harder questions let's go to the hard
problem of consciousness yeah is this an
important problem for us to think about
and solve on the engineering of
intelligence side of your work of our
dream you know it's unclear so you know
again this is a deep problem part
because it's very difficult to define
consciousness and
and there is the debate among
neuroscientist and about whether
consciousness and philosophers of course
whether consciousness is something that
requires flesh and blood so to speak yes
or could be you know that we could have
silicon devices that are conscious or up
to statement like everything has some
degree of consciousness and some more
than others this is like Giulio Tononi
and she would just recently talk to
Christophe Koch okay so he a crystal
force my first graduate student yeah do
you think it's important to illuminate
aspects of consciousness in order to
engineer intelligence systems do you
think an intelligent system would
ultimately have consciousness are they
to the interlinked you know most of the
people working in artificial
intelligence I think who'd answer
we don't strictly need the consciousness
to have an intelligent system that's
sort of the easier question because yeah
because it's it's a very engineering
answer to the question yes that's the
Turing test will run in consciousness
but if you were to go do you think it's
possible that we need to have so that
kind of self-awareness we may yes so for
instance I I personally think that when
test a machine or a person in a Turing
test in an extended to interesting I
think consciousness is part of what we
require in that test you know in
priestly to say that this is intelligent
Christophe disagrees so as he does yeah
it despite many other romantic notions
he who
he disagrees with that one yes that's
right
so you know we would see do you think as
a quick question
Ernest Becker fear of death
do you think mortality and those kinds
of things are important for well for
consciousness and for intelligence the
finiteness of life finiteness of
existence or is that just the side
effect of evolutionary side effect is
useful to a for natural selection do you
think this kind of thing that we're
gonna this interview is gonna run out of
time soon our life will run out of time
soon do you think that's needed to make
this conversation good and in life good
you know I never thought about it is it
a very interesting question I think
Steve Jobs in his commencement speech at
Stanford argued that you know having a
finite life was important for for
stimulating achievement so I was a
different yeah I live every day like
it's your last right yeah yeah so I
rationally I don't think strictly you
need mortality for consciousness but oh
no they seem to go together in our
biological system yeah you've mentioned
before and students are associated with
alpha go immobilize the big recent
success stories in the eye and I think
it's captivated the entire world of what
I can do so what do you think will be
the next breakthrough and what's your
intuition about the next breakthrough of
course I don't know where the next
breakthroughs is I think that there is a
good chance as I said before that the
next breakthrough would also be inspired
by you know neuroscience
but which one I don't know
and there's so MIT has this quest for
intelligence you know and there's a few
moon shots which in that spirit which
ones are you excited about what which
projects kind of well of course I'm
excited about one of the moon shots with
it which is our Center for brains minds
and machines history the one which is
filip fully funded by NSF and it's a it
is about visual intelligence it's an
area that one has a particularly about
understanding visual intelligence or
visual cortex and and visual
intelligence in the sense of how we look
around ourselves and understand the word
around ourselves you know meaning what
what is going on how we could go from
here to there without hitting obstacles
you know whether there are other agents
people in the market these are all
things that we perceive very quickly and
and it's something actually quite close
to being conscious not quite but now
there is this interesting experiment
that was run at Google X which is in a
sense is just a virtual reality
experiment but in which they had subject
sitting in a chair with goggles like
oculus and so on
earphones and they were seeing through
the eyes of a robot nearby two cameras
microphones for a/c mossad their sensory
system was there and the impression of
all the subject very strong they could
not shake it off was that they were
where the robot was they could look at
themselves from the robot and still feel
they were they were where the robot is
they were looking their body their self
were had moved so some aspect of scene
understanding has to have ability to
place yourself have a self-awareness
about your position in the world and
what the world is right so yeah so we
may have to solve the hard problem of
consciousness on their way yes but it's
quite quite quite a moonshot eyes so if
you've been an adviser to some
incredible minds including demis
hassabis Christophe Co I'm not sure like
you said all went on to become seminal
figures in their respective fields from
your own success as a researcher and
from perspective as a mentor of these
researchers having guided them Madhvi so
what does it take to be successful in
science and engineering careers whether
you're talking to somebody in their
teens 20s and 30s what does that path
look like it's curiosity and having fun
and I think is important also having fun
with other curious minds it's the the
people you surround with - so yeah fun
and curiosity is there mentioned Steve
Jobs is there also an underlying
ambition that's unique that you saw or
is it really does boil down to
insatiable curiosity and fun well of
course you know it's been cured
using active and ambitious way yes
definitely but I think sometime in in
science there are friends of mine who
are like this you know there are some of
the scientists like to work by
themselves and kind of communicate only
when they complete their work or
discover something I think I always
found the the actual process of you know
discovering something is more fun if
it's together with other intelligent and
curious and fun people so if you see the
fun in that process of the side effect
of that process will be the election of
discovering something yes so as you've
led many incredible efforts here what's
the secret to being a good advisor
mentor leader in a research setting is
that similar spirit or yeah what what
advice could you give to people young
faculty and so on it's partly repeating
what I said about an environment that
should be friendly and fun and ambitious
and you know I think I learned a lot
from some of my advisers and friends and
some of our physicists and there was
reason this behavior that was encouraged
of when somebody comes with a new idea
in the group you're unless is really
stupid but you are always enthusiastic
and then and the other two just for a
few minutes for a few hours then you
start you know asking critically a few
questions testing but you know this is a
process that is I think it's very very
good this you have to be enthusiasm time
people are very critical from
beginning that's that's that's not yes
you have to give it a chance yes let's
see to grow that said with some of your
ideas which are quite revolutionary so
there's a witness especially in the
human vision side and neuroscience side
there could be some pretty heated
arguments do you enjoy these dessert a
part of science and I could academic
pursue see you enjoy yeah is it is that
something that happens in your group as
well yeah absolutely I also spent some
time in Germany again that is this
tradition in which people are more
forthright less kind than here so you
know in the u.s. you when you write a
bad letter you still say this guy's nice
yes so yet here in America its degrees
of nice yes it's all just degrees of
Nicaea right right so as long as this
does not become personal and it's really
like you know a football game with these
rules that's great so if you somehow
found yourself in a position to ask one
question of an Oracle like a genie maybe
a god whoa and you're guaranteed to get
a clear answer what kind of question
would you ask what what would be the
question you would ask in the spirit of
our discussion it could be how could be
how could I become ten times more
intelligent and so but see you only get
a clear short answer so do you think
there's a clear short answer to that no
and that's the answer you'll get yeah
okay so you've mentioned flowers of
Algren odd oh yeah this is a story that
inspires you in your childhood as this
story of a mouse and human achieving
genius-level intelligence and then
understanding what was happening while
slowly becoming not intelligent again in
this tragedy of
intelligence and losing intelligence do
you think in that spirit and that story
do you think intelligence is a gift or
curse from the perspective of happiness
and meaning of life you try to create
intelligence system that understands the
universe but at an individual level the
meaning of life do you think
intelligence is a gift it's a good
question
I don't know as one of this as one
people consider the smartest people in
the world in some in some dimension at
the very least what do you think no no
it may be invariant to intelligence
likely of happiness would be nice if it
were that's the hope
yeah you could be smart and happy and
clueless unhappy yeah as always on the
discussion of the meaning of life it's
probably a good place to end Tommaso
thank you so much for talking today
thank you this was great
you