Australian Sky & Telescope - April 2018

10 AUSTRALIAN SKY & TELESCOPE April 2018

BOYAJIAN’S STAR: NASA / JPL-CALTECH; KEPLER-90: NASA / WENDY STENZEL

T Kepler-90i is the third planet from the left in

this artist’s illustration of its scrunched system.

All eight planets would fit within Earth’s orbit.

NEW DATA ON KIC 8462852, also

known as Tabby’s Star, show that dust —

not an alien megastructure — is causing

its mysterious behaviour. Results will

appear in the Astrophysical Journal

Letters.

Tabetha Boyajian (Louisiana State

University) and colleagues reported

the star’s discovery in 2016 in archived

Kepler data. The star dipped in

brightness by up to 22% for several days

New observations of Tabby’s Star

A UNIQUE COLLABORATION between

an astronomer and a Google engineer

has led to the creation of a kind of

artificial intelligence known as a neural

network. Set to work on data from

the Kepler mission, the algorithm

discovered two new Earth-size planets,

one of which is part of the first eight-

planet system outside our own.

Neural networks are loosely based

on connections between neurons in

the brain. Each ‘neuron’ is a simple

mathematical formula that acts like a

switch, turning on when it recognises

a certain pattern. Each set of neurons

passes its results on to the next layer.

By combining layers, Christopher

Shallue (Google Brain) built an

Neural network finds two Earth-Size planets

algorithm that can identify exoplanet

transit signatures in stellar light curves.

It learned how by training on a set of

15,000 Kepler light curves previously

studied by humans. Andrew Vanderburg

(University of Texas, Austin) then ran a

preliminary sample of 670 stars through

the standard pipeline to identify signals

to feed to the algorithm. The neural net

caught two rocky planets, Kepler-90i

and Kepler-80g, both having roughly

14-day orbits. The discovery appears in

the January 18 issue of the Astronomical

Journal.

Kepler-90i is the eighth planet

discovered in its tightly packed system.

Kepler-80g is the sixth planet in its

system, orbiting its star in resonance

with four of its companions. The

resonance keeps the orbits stable,

even though they’re more compressed

than Kepler-90’s. Neither planet

offers a pleasant place to live: Surface

temperatures are 436°C and 147°C,

respectively.

Eventually, Shallue and Vanderburg

want to expand their project to study

all 150,000 stars in the Kepler field, but

they’ll proceed with caution. A neural

network is only as good as it was taught

to be; it can’t generalise what it has

learned in the same way humans can.

■ MONICA YOUNG

NEWS NOTES

at a time. Long-term observations also

showed the star fading over the years.

To obtain round-the-clock

observations that could catch the star

as it was dipping, Boyajian launched a

Kickstarter campaign to secure funding

for observations with the Las Cumbres

Observatory (LCO), a network of 0.4-m

robotic telescopes around the globe.

In May 2017 the team caught the

first real-time change, a minor event

that obscured 1% of the star’s light. The

star’s blue light dimmed slightly more

than redder wavelengths during the

event, indicating that whatever material

is involved, it’s not solid. That’s a strike

against the unlikely idea of ‘alien

megastructures’ around the star.

Instead, the data suggest cosmic dust,

1,000 times smaller than the smallest

grains of sand, is the culprit. Since the

star’s light exerts radiation pressure that

would easily blow such small grains out

of orbit, any dust near the star must

have arrived there recently, perhaps

from a swarm of comets, asteroids or a

dusty planetesimal. Dust that has been

ejected to greater distances might dim

the star over longer time scales.

Most dust-producing objects would

orbit the star. Based on the estimated

orbit, Boyajian and colleagues have

predicted that the next set of dips

should occur in June 2019.

■ JOHN BOCHANSKI