Weird Science
Robots could help us find aliens
If we can’t find other life in the universe, machines may be able to do it for us.
A study from the UK’s Plymouth University drew on artificial neural networks (ANNs), to classify planets into five types. These would estimate probability of life, which could be used in future interstellar exploration missions.
ANNs attempt to replicate the way the human brain learns and are one of the main tools used in machine learning.
They’re particularly good at identifying patterns that are too complex for a biological brain to process.
The study team trained its network to bunch planets into five different categories based on whether they are most like present-day Earth, early Earth, Mars, Venus or Saturn’s moon Titan.
All five of these objects are rocky bodies known to have atmospheres, and are among the most potentially habitable objects in our solar system.
“We’re interested in these ANNs for prioritising exploration for a hypothetical, intelligent, interstellar spacecraft scanning an exoplanet system at range,” says study leader Christopher Bishop.
Atmospheric observations, known as spectra, of the five solar system bodies are presented as inputs to the network, which is then asked to classify them in terms of the planetary type.
As life is currently known only to exist on Earth, the classification uses a “probability of life” metric, which is based on the relatively well-understood atmospheric and orbital properties of the five target types.
Bishop has trained the network with over a hundred different spectral profiles, each with several hundred parameters that contribute to habitability.
So far, the network performs well when presented with a test spectral profile that it hasn’t seen before.
“Given the results so far, this method may prove to be extremely useful for categorising different types of exoplanets using results from ground-based and nearEarth observatories,” says Dr Angelo Cangelosi, the supervisor of the project.
The technique may also be suited to selecting targets for future observations, given the increase in spectral detail expected from upcoming space missions such as Esa’s Ariel Space Mission and Nasa’s James Webb Space Telescope.
Why we struggle with g
Despite seeing it millions of times in picture books, novels, newspapers and emails, people are essentially unaware of the more common version of the lower-case print letter g.
Most people don’t even know that two forms of the letter — one usually handwritten, the other typeset — exist.
And if they do, they can’t write the typeset one. They can’t even pick the correct version out of a line-up.
“We think that if we look at something enough, especially if we have to pay attention to its shape as we do during reading, then we would know what it looks like, but our results suggest that’s not always the case,” says Michael McCloskey, a cognitive scientist at Johns Hopkins University in the US, and coauthor of a new study.
“What we think may be happening is that we learn the shapes of most letters because we have to write them in school.
“Looptail g is something we’re never taught to write, so we may not learn its shape.”
Unlike most letters, g has two lowercase print versions.
There’s the opentail one that almost everyone uses when writing by hand; it looks like a loop with a fishhook hanging from it.
Then there’s the looptail g, by far the more common, seen in everyday fonts such as Times New Roman and Calibri and, hence, in most printed and typed material.
To test people’s awareness of the g they write and the g they tend to read, the researchers conducted a three-part experiment.
First, they wanted to find out whether people knew there were two lower-case print gs. They asked 38 adults to list letters with two lower-case print varieties. Two named g. Only one could write both forms correctly.
Next, the researchers asked 16 new participants to silently read a paragraph filled with looptail gs, but to say each word with a g aloud.
Immediately after participants finished, having paid particular attention to each of 14 gs, they were asked to write the g that they just saw 14 times.
Half of them wrote the wrong type, the opentail. The others attempted to write a looptail version, but only one could.
Finally, the team asked 25 participants to identify the correct looptail g in a multiple-choice test with four options. Just seven succeeded.
“They don’t entirely know what this letter looks like, even though they can read it,” says study coauthor Gali Ellenblum, a graduate student in cognitive science.
“This is not true of letters in general.”
This outlier g seems to demonstrate that our knowledge of letters suffers when we don’t write them. And as we write less and become more dependent on electronic devices, the researchers wonder about the implications for reading.
“What about children who are just learning to read? Do they have a little bit more trouble with this form of g because they haven’t been forced to pay attention to it and write it?” McCloskey says.
“That’s something we don’t really know. Our findings give us an intriguing way of looking at questions about the importance of writing for reading.”