Blaise Agüera y Arcas: How computers are learning to be creative
בלייז אגוארה ארקס: איך מחשבים לומדים להיות יצירתיים
Blaise Agüera y Arcas works on machine learning at Google. Previously a Distinguished Engineer at Microsoft, he has worked on augmented reality, mapping, wearable computing and natural user interfaces. Full bio
Double-click the English transcript below to play the video.
that works on machine intelligence;
שעובד על למידה חישובית;
of making computers and devices
למחשבים ומכשירים
that brains do.
interested in real brains
in the things that our brains do
to the performance of computers.
has been perception,
הוא תפיסה,
out there in the world --
for example, that our team makes,
שהצוות שלנו מוציא,
on Google Photos to become searchable,
בגוגל פוטוז להיות ניתנות-חיפוש,
out there into the world.
our work on machine perception
עבודתנו על תפיסה חישובית
with the world of machine creativity
לעולם היצירתיות החישובית,
had a penetrating insight
between perception and creativity.
has a statue inside of it,
is to discover it."
Michelangelo was getting at
is an act of imagination
and perceiving and imagining,
with a brief bit of history
the heart or the intestines,
about a brain by just looking at it,
פשוט מלהסתכל עליו,
of this thing all kinds of fanciful names,
בכל מיני שמות דמיוניים,
doesn't tell us very much
developed some kind of insight
איזה מין של הבנה
Santiago Ramón y Cajal,
סנטיאגו רמון אי קחאל,
or render in very high contrast
their morphologies.
that he made of neurons
of different sorts of cells,
של תאים מסוגים שונים,
was quite new at this point.
הייתה חדשה למדי בעת הזו
very, very long distances --
to some people in the 19th century;
במאה התשע-עשרה;
were just getting underway.
of Ramón y Cajal's, like this one,
של רמון אי קחאל, כגון זה,
that Ramón y Cajal started.
ממשתפי הפעולה שלנו
of Neuroscience.
is about one cubic millimeter in size,
בגודלה למילימטר מעוקב אחד,
very small piece of it here.
tiny block of tissue.
of hair is about 100 microns.
much, much smaller
electron microscopy slices,
מוגדלות מיקרוסקופ-אלקטרוני,
in 3D of neurons that look like these.
של תאי עצב, הנראים ככה.
style as Ramón y Cajal.
בקירוב של רמון אי קחאל.
be able to see anything here.
פה שום דבר.
one neuron to another.
ahead of his time,
over the next few decades.
was advanced enough
הטכנולוגיה שלנו התקדמה דיה
experiments on live neurons
אמיתיים בתאי עצב חיים
when computers were being invented,
of modeling the brain --
as Alan Turing called it,
כמו שאלן טיורינג כינה אותו,
looked at Ramón y Cajal's drawing
הביטו בציורו של רמון אי קחאל
imagery that comes from the eye.
דימויים המגיעים מהעין.
like a circuit diagram.
דיאגרמת מעגל חשמלי.
in McCulloch and Pitts's circuit diagram
המעגל של מקולוק ופיטס
of computational elements
כסדרה של רכיבים חישוביים
one to the next in a cascade,
visual information would need to do.
for us to do with our brains.
that for a computer,
just a few years ago.
this task is easy to do.
and the word "bird,"
connected to each other
בין מערכת תאי עצב מקושרים
inside our visual cortices,
בתוך קליפתנו הראייתית,
to have the capability
on the computer.
that actually looks like.
about as a first layer of neurons,
כשכבה הראשונה של ניורונים,
how it works in the eye --
after another layer of neurons,
אחרי הבאה של ניורונים,
of different weights.
of all of those synapses.
של כל הסינפסות האלה.
properties of this network.
של הרשת הזו.
or a small group of neurons
those three things --
את שלושת הדברים האלה --
in the neural network,
these synapses in the neural network.
ברשתות העצביות.
is just a simple formula,
going on there, of course,
of mathematical operations.
that if you have one equation,
by knowing the other two things.
אם יודעים את שני הדברים האחרים.
that the picture of a bird is a bird,
and w and x are known.
you know the pixels.
אתם יודעים את הפיקסלים.
a relatively straightforward problem.
זו למעשה בעיה די פשוטה.
and you're done.
doing exactly that.
on a mobile phone,
amazing in its own right,
billions and trillions of operations
כל כך הרבה מליארדים וטריליונים של פעולות
picture of a bird,
"Yes, it's a bird,"
with a network of this sort.
של הציפור עם רשת מסוג כזה.
and the y is the unknown.
difficult part, of course,
do we figure out the w,
of solving for w,
with the simple equation
איך לעשות את זה: 6 = 2 *w,
it's the inverse to multiplication,
very non-linear operation;
מאוד לא לינארית;
to solve the equation
is fairly straightforward.
a little algebra trick,
בואו נשחק בטריק אלגבראי פשוט,
to the right-hand side of the equation.
about it as an error.
for w the right way,
to minimize the error,
כדי להקטין את השגיאה,
computers are very good at.
sort of play Marco Polo,
successive approximations to w.
הוא מקבל קרובים עוקבים ל W.
but after about a dozen steps,
אבל אחרי בערך שנים עשר צעדים,
which is close enough.
a lot of known x's and known y's
through an iterative process.
that we do our own learning.
את הלמידה בעצמנו.
this is not a bird."
for those neural connections.
אנחנו פותרים עבור החיבורים העצביים האלה.
x and w fixed to solve for y;
כדי לפתור עבור Y;
Alex Mordvintsev, on our team,
אלכס מורדבינטסב, בצוות שלנו,
with what happens if we try solving for x,
אם אנחנו מנסים לפתור עבור X,
that you've trained on birds,
שאימנתם על ציפורים,
the same error-minimization procedure,
באותו תהליך מזעור שגיאות,
trained to recognize birds,
generated entirely by a neural network
שמייוצרת לגמרי על ידי רשת עצבית
rather than solving for y,
by Mike Tyka in our group,
על ידי מייק טייקה בקבוצה שלנו,
of William Kentridge's artworks,
של ווליאם קמטרידג',
over the space of different animals,
במרחב של חיות שונות,
to recognize and distinguish
morph from one animal to another.
כמו של אשר, מחיה אחת לאחרת.
have tried reducing
out of the space of all things
over that entire surface,
you make a kind of map --
אתם עושים סוג של מפה --
the network knows how to recognize.
שהרשת יודעת איך להכיר.
"armadillo" is right in that spot.
בדיוק בנקודה הזו בנקודה ההיא.
of networks as well.
גם עם סוגים אחרים של רשתות.
to recognize faces,
in a y that says, "me,"
psychedelic picture of me
ופסיכדלית שלי
multiple points of view at once
של נקודות מבט יחד
to get rid of the ambiguity
or another pose,
another kind of lighting.
או סוג אחר של תאורה.
this sort of reconstruction,
בסוג מסווים של תמונת הנחייה
of different points of view,
מנקודות מבט שונות,
his own face as a guide image
משתמש בפנים של עצמו כתמונה מנחה
to reconstruct my own face.
כדי לבנות מחדש את הפנים שלי.
that optimization process.
more like a coherent face,
יותר כמו פנים ברורות,
with a blank canvas
that is itself already some other image.
שהוא בעצמו כבר תמונה אחרת.
that is designed to categorize
man-made structures, animals ...
מבנים מעשה ידי אדם, חיות...
with just a picture of clouds,
what it sees in the clouds.
מה היא רואה בעננים.
you spend looking at this,
will see in the clouds.
to hallucinate into this,
ברשת הפנים כדי להזות לתוך זה,
zooms hallucinates, zooms.
of the network, I suppose,
של פוגה של הרשת, אני מניח,
is eating its own tail.
What do I think I see next?"
מה אני חושב שאני רואה עכשיו?"
called "Higher Education" --
שנקראה "חינוך גבוה יותר" --
marijuana was legalized.
is not constrained.
because they're really fun to look at.
בגלל שבאמת כיף להביט בהן.
a camera that takes a picture,
writes a poem using neural networks,
בשימוש ברשתות עצביות,
has been trained
are very intimately connected.
things in the world,
Michelangelo really did see
any being, any alien
perceptual acts of that sort
machinery that's used in both cases.
שבשימוש בשני המקרים.
and creativity are by no means
הן בשום צורה
that can do exactly these sorts of things.
שיכולים לעשות בדיוק דברים מהסוג הזה.
the brain is computational.
in designing intelligent machinery.
of those early pioneers,
is not just about accounting
we modeled them after our minds.
to understand our own minds better
להבין את המוחות שלנו טוב יותר
ABOUT THE SPEAKER
Blaise Agüera y Arcas - Software architectBlaise Agüera y Arcas works on machine learning at Google. Previously a Distinguished Engineer at Microsoft, he has worked on augmented reality, mapping, wearable computing and natural user interfaces.
Why you should listen
Blaise Agüera y Arcas is principal scientist at Google, where he leads a team working on machine intelligence for mobile devices. His group works extensively with deep neural nets for machine perception and distributed learning, and it also investigates so-called "connectomics" research, assessing maps of connections within the brain.
Agüera y Arcas' background is as multidimensional as the visions he helps create. In the 1990s, he authored patents on both video compression and 3D visualization techniques, and in 2001, he made an influential computational discovery that cast doubt on Gutenberg's role as the father of movable type.
He also created Seadragon (acquired by Microsoft in 2006), the visualization technology that gives Photosynth its amazingly smooth digital rendering and zoom capabilities. Photosynth itself is a vastly powerful piece of software capable of taking a wide variety of images, analyzing them for similarities, and grafting them together into an interactive three-dimensional space. This seamless patchwork of images can be viewed via multiple angles and magnifications, allowing us to look around corners or “fly” in for a (much) closer look. Simply put, it could utterly transform the way we experience digital images.
He joined Microsoft when Seadragon was acquired by Live Labs in 2006. Shortly after the acquisition of Seadragon, Agüera y Arcas directed his team in a collaboration with Microsoft Research and the University of Washington, leading to the first public previews of Photosynth several months later. His TED Talk on Seadragon and Photosynth in 2007 is rated one of TED's "most jaw-dropping." He returned to TED in 2010 to demo Bing’s augmented reality maps.
Fun fact: According to the author, Agüera y Arcas is the inspiration for the character Elgin in the 2012 best-selling novel Where'd You Go, Bernadette?
Blaise Agüera y Arcas | Speaker | TED.com