The best planet hunter is the human eye: Sergio Álvarez Leiva at TEDxMadrid


Translator: Javi Garriz
Reviewer: antonella broglia Kepler said something along the lines, that the diversity of things
that nature hides and the amount of treasures
we can find in outer space, do nothing but feed our minds with certain curiosity. In 2009, NASA started
the «Kepler Mission», I think Carlos may have mentioned it
early this morning. The goal of the Kepler Mission
is locating planets similar to Earth outside the solar system,
in a small portion of the universe with the right conditions, using a relatively simple method
of detection. Thanks to the Kepler Mission,
we have found 151 planets to date. And there are more than 3,588
candidates waiting for the scientific committee
to declare them confirmed planets. What the relatively small Kepler
satellite does, is to record the brightness of the stars and find things that cross their path
right in front of them. The satellite detects the planets
when they are passing over the stars. I’m going to explain it with my hands. This would be a star and the Kepler is recording every data on its brightness and what the Kepler does
is to go find all these little eclipses caused by the bodies
that are passing over them. When this eclipse occurs,
the light of the stars varies a little, it minimally diminishes. These are the variations that Kepler
tries to detect with powerful lenses. Kepler sends the data to NASA,
where they analize them and try to find the planets. This is the brightness data
of a random star. That is what it looks like. These are 90 days of light
of a single star from this portion of the universe. But as you all know, there are thousands, millions,
or probably infinite stars in the universe. So there is more than one graphic. NASA computers have to analize
many graphics, each one unique, and they have to look for certain patterns in order to find these planets. I am here to tell you about a bet
we made some time ago. We said, “We believe humans
are still better than machines doing certain things”. And one of them is to recognize patterns. One of the innate abilities
of the human mind is to recognize such patterns. We can teach machines
how to recognize patterns. We can give them a set of probabilities
and rules that regulate these patterns, but in the case of the universe,
it’s extremely complicated. What we had to do was relatively simple. We are going to offer the data
to the people, NASA will give away the data. We are going to give them the data
and tell them, “Look for these variations
in the graphics”. We are going to explain what they are and they have to mark them. Instead of giving them to one person, we give them to many people
at the same time in such a way that,
if many variations appear in the same part of the graph, we would consider it
a correct observation, and if there were discussion,
we would discard that data. We had to build an interface,
a web application, a place where people who had
little knowledge about the topic could go, analize data,
find the more evident patterns, and have fun, get hooked somehow. But also, we had to do something
that would allow the experts to analize the data in depth,
to make sense of that sea of dots, and detect the most complicated transits. That is how Planet Hunters was born. It is a website, a joint work along with
Oxford University and Zooniverse, which is an organization that makes
citizen science projects like this one, where each of you can enter and start looking for planets,
with just two mouse clicks. As of today, there have been more than
16,600,000 observations. Which means that these graphics
have been analyzed more than 16 million times. This number sounds impressive. A lot of people are investing
a lot of time in this. Instead of playing coloured fruits
and creating virtual animals, people are doing these things that can make this world a little better. Thanks to all this people,
we have discovered two planets. They are two real planets: they have a name and a body; “Planet Hunters 1”
and “Planet Hunters 2”. PH1 is a planet six times
the size of Earth approximately, and it takes 183 days to complete
its orbit. What is interesting about PH1 is that
it is located in a galaxy of four stars, which is something more
than extraordinary. It orbits around what is called
“binary stars”. These are two stars that orbit each other. PH2 is ever bigger,
10 times the size of the Earth and its full orbit takes 282 days. If we had to find the Earth
using this method we would have needed to analyze
years worth of data to see how the Earth completes its orbit. To validate a candidate,
we need those transits to repeat themselves at least 3 times
in all the data we have. On top of the two planets,
there are 34 candidates more. The most important thing
in this project is that the data have already been analyzed by machines,
by computers. And it took us humans to find what they were missing. I remember a line I couldn’t help
myself to write down, I was not able to learn it by heart,
I’m afraid, that someone said on the Internet: “scientists are people who study
nature in its essence they look for patterns in the randomness
so that they can, transform the chaos
in order and understanding”. This is what all that people made
with Planet Hunters. We offered them a bunch of data,
a huge chaos of graphics and dots, and they have searched,
they invested time, patiently, very thoroughly. They even got together in houses
to analyze a graphic together, and we achieved for them
to make real science. They have been making science
from their homes, from their cars, from the beach, from their favorite café, just by having Internet access
and clicking a few times. With this and since it’s raining I’ll get to the end. Two ideas, we were talking about
unfinished things, I want to leave you with two more. The first one is space. Probably the mission of discovering
and analyzing space, will always remain unfinished, among other things because Kepler,
in July 2012, got a failure in one of its four engines, and in May this year
a second engine failed, and on the 8th of August the Kepler
mission was shut down. The satellite doesn’t send
data to Earth anymore, but all the data it has been sending can still be analyzed on the website. The second idea is that we proved again that we are still better than machines. How far would we be able to go
with machines in order for us not to ever be
better than them? I don’t know but I’ll be delighted
to be here to see it. I didn’t want to leave without thanking
all the users that invested their time
in making real science, that helped discover these two planets; to Kepler for planting the seed
of curiosity, and you for letting me share this project that for me is probably the most important
in my career. Thank you very much. (Applause)

2 Comments

Add a Comment

Your email address will not be published. Required fields are marked *